Coyote Blog

Many media outlets were calling the post-Harvey flooding in East Texas a "1000-year" flood.  Forget for a moment on the craziness of saying this is a 1000-year flood when we have at most about 150 years of weather records for the Houston area.  Consider something else -- that our intuition about outlier events tends to suck.

Let's say the flood affected a quarter of Texas.  This is probably an exaggeration, but it will help the following analysis be conservative.   Based on numbers from Wikipedia, Texas has 268,597 square miles of land area and the whole globe has 57.5 million square miles of land area.  This means that a quarter of Texas is about 1/1000 of the land area of the globe.  Even if this were truly a 1000-year storm, we should see such a storm over a similar area of land every single year on average somewhere in the world.  And if you add in other weather events that I have seen described as "thousand year", including snowfalls, heat waves, cold waves, droughts, etc. then we should be seeing a thousand-year weather event of some sort over a similar area as that affected by Harvey every few months.

A while back I wrote about the Obama Administration's near exponential expansion of EEO reporting

 It takes the current EEO-1 (the annual exercise where we strive for a post-racial society by racially categorizing all of our employees) and makes it something like 15-20 times longer.  In addition, rather than simply "count" an employee as being on staff in a certain race-gender category, we now have to report their income and hours worked.  Either I will have to hire staff just to do this stupid report, or I will again (like with Obamacare) have to pay a third party thousands of dollars a year to satisfy yet another government reporting requirement.  This is utter madness.

Get this -- the report has 3600 individual cells that must be filled in.  And this is in addition to the current EEO-1 form, which also still has to be filled out.  The draft rule assumes 6-7 hours per company per year for this reporting.  They must be joking.

Fortunately, the Trump Administration has at least temporarily suspended this requirement:

On Tuesday, the White House suspended a burdensome reporting requirement for employers that would have cost them $400 million while yielding information of questionable value. It did so in rejecting changes to the EEO-1 form made at the end of the Obama administration.

The White House Office of Management and Budget stated that the pay collection and reporting requirements “lack practical utility, are unnecessarily burdensome, and do not adequately address privacy and confidentiality issues.” It explained its reasoning in a letter to the Chair of the Equal Employment Opportunity Commission, Victoria Lipnic.

The Obama administration had claimed that rewriting the form to include 3,660 boxes for companies to check or fill out would help identify wage discrimination. But very little of the information it sought would have shed any light on potential wage discrimination.

First, I want to congratulate @apple for introducing a $1000 phone with features like wireless charging and an edge-to-edge screen that my last two or three android phones have already had.  Perhaps the most, or only, interesting new feature is the facial recognition.  Apple is abandoning fingerprint scanning in favor of facial recognition to unlock the phone.

I mention the law enforcement angle in the title because it has been a bone of contention how far law enforcement can go to make someone unlock their phone.  Clearly, when unlocking was PIN only, one only had to declare they forgot and no one could really disprove that.  With fingerprint scanning, it has been a point that is still in the courts (I believe) as to whether LE can force someone to unlock the phone with their finger.  Now, however, all they will have to do is hold the phone up to the suspect's face.  This less invasive unlocking technique is probably an everyday hassle reduction, but will make the phone incrementally less secure from snooping.

Morbid postscript:  I wonder if this works on a corpse?  Is there a heat sensor of some sort, or are the kids going to be saying "let's get the eyes on dad's body open so we can get his phone unlocked".

One of the best innovations on the web has been customer review scores.  I use the reviews of products at Amazon.com, Tripadvisor, Yelp, and Opentable all the time to aid in my buying.  Sure they can be frustrating -- some reviewers will petulantly give 1 star reviews for absurd issues or failings.  And I know that as much as reviews on Tripadvisor, Google Places, and Facebook can drive me crazy, they help me improve my business.

But these systems only work when they are run with integrity. I once had to get a Tripadvisor review deleted because it was fraudulent (made up claims from a disgruntled employee rather than a customer).  It was a long, uphill battle to get that one review deleted, as it should be.

Unfortunately, VRBO and HomeAway (I think they are the same company now) have abandoned this integrity.  For those that do not know, these sites feature rental of vacation homes and apartments.  We love this travel option - often we can get a nice 2 bedroom condo with kitchen and living room for the same price as a hotel room.  On this site there are often hundreds or thousands of options for rentals, and so customer reviews can be an important source of information in choosing.  Does it really look like its pictures?  Was everything there that was promised?  Are there any location or noise issues?  Essentially, reviews make sure the landlord cannot try to hide issues from travelers.

It used to be you could just log in and review the location, just like one does with a product on Amazon.  I think there was some testing to make sure you had actually rented it, but this is easy and Amazon has the same thing where it tags reviews with something like "confirmed buyer" or whatever.  But VRBO has now gone to a system where the landlord can essentially opt out of the review process.  If they don't send you a review link, you can't review.  In other words, you can't review without the owners permission.  And, as you may guess, owners with properties that have flaws that would readily be pointed out by reviewers do not allow one to review.

To compound the problem, VRBO hides all this.  For example, we rented this flawed beach home in San Diego.  It was wonderful in every way except for one -- the properties below and around it seem to be preferred destinations for loud groups of frat boys partying.  We pretty much got no sleep.   I wanted to warn future customers of this potential issue, but that is impossible because the landlord will not send me a VRBO review link, and that is the only way I can review it.  VRBO hides this because the listing says "This property doesn't have any reviews yet!"  That sounds far more innocent than the more accurate statement, which would be "This property does not choose to participate in the review process."

A few weeks ago I wrote:

If you want to convince me of the need for restrictions on any substances, such as narcotics, you have to convince me of three things:

1. That incarcerating users is somehow better for them than their addiction
2. That ethically abusers of the substance are more worthy of our attention and intervention than legitimate users who benefit from the substance and whose access will likely be restricted
3. That the negative social costs of the substance's use are higher than the inevitable social costs of the criminal black market (including the freedom-reducing policing laws implemented in response) that will emerge when its use or purchase is banned

Think in particular about point #2 when reading this:

Arizona would limit all initial opioid prescriptions to five days for new patients under sweeping guidelines recommended Wednesday by Gov. Doug Ducey's administration.

The plan also would limit maximum doses for pain medication, implement steps to taper down pain medications and require pain prescriptions to be filed electronically, rather than on paper, to limit diversion of drugs.

Consider that many legitimate users will need more than the legal maximum dosage to control their pain, and thus the issue becomes whether we want to essentially torture innocent sick people by forcing them to remain in excruciating pain in exchange for (possibly) reducing the number of accidental deaths from abusers of these drugs (I say possibly because over the last 40 years the government war on drugs has had such a super stellar track record in reducing narcotic usage).

To me the answer to this tradeoff is obvious but I am willing to admit it is a tradeoff subject to debate.  But the article linked has no debate.  There is not a single mention of any downsides to the rules, or any potential harm to legitimate users.

When I was at consultant McKinsey & Co, one of their philosophies in doing cost reduction studies was that you don't cut staffing without first cutting back the work.  Identify the activities that don't need to be done or can be streamlined, change the processes to match, and then cut staffing.

This, of course, is not the way it usually happens, even in good companies.  Most companies just whack staff counts by some percentage, perhaps across the board and perhaps weighted by intuitions as to where the company is fat.  In a good company with good managers and good incentives, the organization can generally be trusted to cut back on the least useful activities in response to the staff cuts.  But in bad organizations with poor incentives, one has no idea if high value or low value activities are being cut.  And in the government, you can almost be assured that when staff and budgets are cut, low-value activities are preserved while high-value core mission activities are cut.  In my world of public parks, staff cuts almost always lead to preservation of bloated headquarters staff while maintenance budgets and staff actually service visitors in parks is slashed.

And then there is this on new rules being imposed by the Trump Administration on NEPA.   If you want to know why infrastructure projects almost never get started and public lands are seldom improved, NEPA is a big part of the reason.  It is something that is desperately in need of reform.  But the Trump Administration appears to be making the same mistake I discussed above, cutting resources without cutting the work required:

Yesterday Greenwire ran a story about how one of the new political appointees at the Department of the Interior issued a memo requiring that National Environmental Policy Act (NEPA) studies (those Environmental Impact Statements you hear so much about) be completed in one year, and be no more than 150 pages long.

If there were ever any doubts that the Trump Administration minions have absolutely no idea what they’re doing, this should put them to rest. Ostensibly intended to “streamline the regulatory process”, blah, blah, blah, its effect will be precisely the opposite; this one memo will delay and stop more projects than anything the environmental community has ever come up with. The activists may be livid, but I promise you that their lawyers are going, “Yee-ha!” I certainly would be if I were still doing NEPA cases.

The NEPA and the hundreds of court decisions interpreting it are painfully clear on how detailed an EIS has to be. Putting artificial and arbitrary limits on an EIS will make it so much easier to show how the EIS does not “take a hard look at the environmental consequences,” contain “a detailed statement of any adverse environmental effects” of a proposed project, etc.

I have been out of town and absolutely consumed by a couple of very weird business issues, so I have not blogged much.  I will try to do a little today, but I will leave you with a quote I liked from Scott Sumner:

Or perhaps (as I've argued elsewhere) there is no such thing as "public opinion". People are like electrons; you can't measure them without changing their positions. ....Most people don't have views that are internally consistent, so their "views" on public policy issues are strongly shaped by the wording of the polls.