Coyote Blog

Of late I have been seeing a lot of examples of people trying to claim that complex, even chaotic multi-variable systems are in fact driven by a single variable.  Whether it be CO2 in climate or government spending in Keynesian views of the economy, this over-simplification seems to be a hubris that is increasingly popular.

The worst example I believe I have ever seen of this was in the editorial page today in the Arizona Republic.  Titled Arizona vs. Massachusetts,  this article purports to blame everything from Arizona's higher number of drunk driving accidents to its higher number of rapes on ... the fact that Arizona has lower taxes.  I kid you not:

In the absence of discernible benefits, higher taxes are indeed a negative. We would all like to keep more of what we earn. That is, if there are not other negative consequences. So, it is reasonable to ask: What do Massachusetts citizens get for these increased public expenditures? A wide range of measures from widely disparate sources provide insight into the hidden costs of a single-minded obsession with lower taxes at all costs.

The results of such an investigation are revealing: Overall, Massachusetts residents earn significantly higher salaries and are less likely to be unemployed than those who live in Arizona. Their homes are less likely to be foreclosed on. Their residents are healthier and are better educated, have a lower risk of being murdered, getting killed in a car accident or getting shot by a firearm than are Arizonans. Perhaps these factors explain the lower suicide rate in Massachusetts than in Arizona as well as the longer life spans.

None of this supposed causation is based on the smallest scrap of evidence, other than the spurious correlation that Arizona has lower taxes at the same time it has more of the bad things the authors don't like.  The authors do not even attempt to explain why, out of the thousands of variables that might have an impact on these disparities, that taxation levels are the key driver, or are even relevant.

Perhaps most importantly, the authors somehow fail to even mention the word demographics.  Now, readers know that I am not very happy with Arizona Conservatives that lament the loss here of the Anglo-Saxon mono-culture.   I think immigration is healthy, and find some of the unique cultures in the state, such as on the large tribal reservations, to make the state more interesting.

However, it is undeniable that these demographic differences create wildly different cultures between Arizona and Massachusetts, and that these differences have an enormous impact on the outcomes the authors describe.  For example, given the large number of new immigrants in this state, many of whom come here poor and unable to speak English, one would expect our state to lag in economic averages and education outcomes when compared to a state populated by daughters of the revolution and the kids of college professors (see immigration data at end of post).  This is made worse by the fact that idiotic US immigration law forces many of these immigrants underground, as it is far harder to earn a good income, get an education, or have access to health care when one does not have legal status.  (This is indeed one area AZ is demonstrably worse than MA, with our Joe-Arpaio-type fixation on harassing illegal immigrants).

By the way, it turns out Arizona actually does pretty well with Hispanic students vs. Massachusetts  -- our high school graduation rate for Hispanics is actually 10 points higher than in MA (our graduation rate for blacks is higher too).  But since both numbers are so far below white students, the heavy mix of Hispanic students brings down Arizona's total average vs. MA.   If you don't understand this issue of how one state can do better than another on many demographic categories but still do worse on average because of a more difficult demographic mix, then you shouldn't be writing on this topic.

Further, the large swaths of this state that are part of various Indian nations complicate the picture.  AZ has by far the largest area under the management of tribal nations in the country -- in fact, almost half the tribal land in the country is in this one state.  These tribal areas typically add a lot of poverty, poor education outcomes, and health issues to the Arizona numbers.  Further, they are plagued with a number of tragic social problems, including alcoholism (with resulting high levels of traffic fatalities) and suicide.  But its unclear how much these are a result of Arizona state policy.   These tribal governments are their own nations with their own laws and social welfare systems, and in general fall under the purview of Federal rather than state authority.  The very real issues faced by their populations have a lot of historical causes that have exactly nothing to do with current AZ state tax policy.

The article engages in a popular sort of pseudo-science.  It drops in a lot of numbers, leaving the impression that the authors have done careful research.  In fact, I count over 50 numbers in the short piece.  The point is to dazzle the typical cognitively-challenged reader into thinking the piece is very scientific, so that its conclusions must be accepted.  But when one shakes off the awe over the statistical density, one realizes that not one of the numbers are relevant to their hypothesis: that the way Arizona runs its government is the driver of these outcome differences.

It's really not even worth going through the rest of this article in detail.  You know the authors are not even trying to be fair when they introduce things like foreclosure rates, which have about zero correlation with taxes or red/blue state models.  I lament all the negative statistics the authors cite, but it is simply insane to somehow equate these differences with the size and intrusiveness of the state.  Certainly I aspire to more intelligent government out of my state, which at times is plagued by yahoos focusing on silly social conservative bugaboos.  I am open to learning from the laboratory of 50 states we have, and hope, for example, that Arizona will start addressing its incarceration problem by decriminalizing drugs as has begun in other states.

The authors did convince me of one thing -- our state university system cannot be very good if it hires professors with this sort of analytical sloppiness.  Which is why I am glad I sent my son to college in Massachusetts.

PS- If the authors really wanted an apples to apples comparison that at least tried to find states somewhat more demographically similar to Arizona, they could have tried comparing AZ to California and Texas.  I would love for them to explain how well the blue state tax heavy model is working in CA.  After all, they tax even more than MA, so things must be even better there, right?  I do think that other states like Texas are better at implementing aspects of the red-state model and do better with education for example.  You won't get any argument from me that the public schools here are not great (though I work with several Charter schools which are fabulous).  For some reason, people in AZ, including upper middle class white families, are less passionate on average about education than folks in other states I have lived.  I am not sure why, but this cultural element is not necessarily fixable by higher taxes.

Update- MA supporters will argue, correctly, that they get a lot of immigration as well.  In fact, numerically, they get about the same number of immigrants as AZ.  But the nature of this immigration is totally different.  MA gets legal immigrants who are highly educated and who come over on corporate or university-sponsored visa programs.  Arizona gets a large number of illegal immigrants who get across the border with a suitcase and no English skills.  The per-person median household income for MA immigrants in 2010 was $16,682 (source).  The per-person median household income for AZ immigrants was $9,716.  35.3% of AZ immigrants did not finish high school, while only 15.4% of MA immigrants have less than a high school degree.  48% of AZ immigrants are estimated to be illegal, while only 19% of those in MA are illegal.  11% of Arizonans self-report that they speak English not at all, vs. just 6.7% for MA (source).

I saw this at Flowing Data -- this is apparently a chart prepared by some sustainability group at MIT to map solar potential of different sites in Cambridge, MA

Look at all the sites marked "excellent".  I have news for the brilliant folks at MIT.  Even the best, flattest roof facing south in Cambridge, MA still rates a "sucks" for solar potential. (source)

Even with massive state and Federal subsidies, those of us who live in the bright red areas find that roof-top solar PV is still an - at best - marginal investment with very long payback times.  We all hope to change this in the future, but there is no way a city like Cambridge with approximately half the solar insolation we get in AZ is going to have "excellent" roof top solar PV sites.

The Town of South Attleboro, MA sent out wildly threatening past due letters for folks with balances as low as 1-cent  (thereby investing at least 42 cents to get one back).  In response to charges that this was stupid, City Collector Debora Marcoccio responded:

A computer automatically printed the letters for any account with a balance remaining, and they were not reviewed by staff before being sent out, Marcoccio said.

"It would be fiscally irresponsible for me to have staff weed through the bills and pull out any below a certain amount," Marcoccio said. " And what would that amount be?"

What, are we living in the 19th century with clerks in a musty room preparing bills by hand?  This fix probably requires one whole entire line of program code in the billing system to fix.  I could probably teach myself to code whatever language the payroll system is written in (my guess is COBOL, which, god help me, I already know) in less time than this woman has spent fielding complaints and media inquiries.  Compare this to what TJIC has to do just to get the mail out.

And don't you love people who don't even have enough spine to make a simple decision about the cutoff for minimum bill size.  I have found this is one of those things the government is really, really bad at -- making decisions under uncertainty  (which covers about all decisions, except routine ones embodied in SOP).  Government has no incentives, in general, for productivity, or production, or customer satisfaction.  The only time government employees get feedback at all is when they get negative feedback from having someone yell at them for making a decision that some higher-up didn't like..  So if a decision is not justifiable either by past precedent/SOP or explicitly by the rules, it is not made.

By the way, I had a personal programming milestone last night.  I finally built a website without using a WYSIWIG editor that formatted the way I wanted it to all in CSS without a single table.  I predict that now that I have finally gotten a decent handle on CSS, which mainly consists of learning all the workarounds for when it doesn't work as you would expect, that someone is about to introduce a whole new system for formatting web pages.

A while back, I criticized a story in the NY Times, as quoted by Kevin Drum, that said that California had among the lowest per capita electricity usage of any state (true) and that this was because of the intelligent regulation regime in the state (yes, but not the way they meant).  The implication of Drum's argument was that there was some sort of efficiency ideal that a smart group of technocrats could reach at limited cost to the state (false). Specifically, Drum argued:

Anyway, it's a good article, and goes to show the kinds of things we
could be doing nationwide if conservative politicians could put their
Chicken Little campaign contributors on hold for a few minutes and take
a look at how it's possible to cut energy use dramatically "” and reduce
our dependence on foreign suppliers "” without ruining the economy. The
energy industry might not like the idea, but the rest of us would.

My response, in part, was this:

Well, here are the eight states in the data set above that the
California CEC shows as having the lowest per capita electricity use:
CA, RI, NY, HI, NH, AK, VT, MA.  All right, now here are the eight
states from the same data set that have the highest electricity prices:  CA, RI, NY, HI, NH, AK, VT, MA.  Woah!  It's the exact same eight states!  The 8 states with the highest prices are the eight states with the lowest per capita consumption.
Unbelievable.  No way that could have an effect, huh?  It must be all
those green building codes in CA.  I suspect Drum is sort of right,
just not in the way he means.  Stupid regulation in each state drives
up prices, which in turn provides incentives for lower demand.  It
achieves the goal, I guess, but very inefficiently.  A straight tax
would be much more efficient.

As part of a presentation I am working on about global warming and proposed California CO2 abatement bill AB52, I had the occasion to do a bit more research.  All of my data is from the Energy Information Administration, whose page URLs keep changing and thus breaking my links but this index page to data seems to stay the same.

I found three factors that seem to be the main drivers of state electricity demand (which is measured in all of the charts below in thousands of kw-h per capita).  The first factor is climate, and certainly California has one of the milder climates.  The chart below looks at residential electricity demand vs. cooling degree days (weighted for population location).  Each data point is a state, with California is shown as the red data point:

We get something similar for heating degree days, with electrical use going down as the climate gets milder, though not as good of a fit, which is not surprising since electricity is less important to heating than cooling.  Since California is well below the line, mild climate can be said to explain some of its lead on other states, but not all.

So I looked next at the percentage of electricity demand that goes to industry.  More heavily industrialized states will have a higher total per capita demand, because heavy industry chews up electricity that other types of businesses do not.  It turns out that California has a relatively low industrial use, which is not surprising given the regulatory environment there and the degree to which industry has been chased out of the state (one would have to be a madman to, all things considered, set up a new factory in California).  So here is the same type of chart of total electrical per capita use by state vs. the % industrial demand, again with each data point a state and California in red:

Again there is a pretty strong relationship, and again we see some but not all of California's low per capita consumption explained.  In effect, states on the left have exported their high-electricity-use industries to the states on the right (or to other countries).

I have saved the most obvious relationship for last:  price.  It turns out unsurprisingly that the states with the highest electricity prices have the lowest per capital consumption:

Rolling climate, industrial intensity, and price together, these factors seem to explain at least 80% of California's efficiency lead over other states.  California government regulatory policy does indeed drive lower electrical consumption, just not exactly the way they would like you to think.  By chasing industries out of the state and raising electricity costs above those of almost every other state, California has reached a lower per capita consumption level.

California is apparently trumpeting its "leadership in energy."  The centerpiece of its claims is its low per capita electricity use.  Arnold is making the claim now, but Kevin Drum was pushing this a while back when he said:

Anyway, it's a good article, and goes to show the kinds of things we
could be doing nationwide if conservative politicians could put their
Chicken Little campaign contributors on hold for a few minutes and take
a look at how it's possible to cut energy use dramatically "” and reduce
our dependence on foreign suppliers "” without ruining the economy. The
energy industry might not like the idea, but the rest of us would.

Max Schulz of the Manhattan Institute is not impressed:

California's proud claim to have kept per-capita energy consumption
flat while growing its economy is less impressive than it seems. The
state has some of the highest energy prices in the country "“ nearly
twice the national average "“ largely because of regulations and
government mandates to use expensive renewable sources of power. As a
result, heavy manufacturing and other energy-intensive industries have
been fleeing the Golden State in droves.

Neither am I.  I addressed this issue a while back in response to Drum's post, but since the meme is going around again, I will excerpt from that old post.

The consumption data is from here.
You can see that there are three components that matter - residential,
commercial, and industrial.  Residential and commercial electricity
consumption may or may not be fairly apples to apples comparable
between states (more in a minute).  Industrial consumption, however, will not be comparable, since the mix of industries will change radically state by state.....

Take two of the higher states on the list.  Wyoming, at the top of
the per capita consumption list, has industrial electricity consumption
as a whopping 58% of total state consumption.  KY, also near the top,
has industrial consumption at 50% of total demand.  The US average is
industrial consumption at 29% of total demand.  CA, NY, and NJ, all
near the bottom of the list in terms of per capital demand, have
industrial use as 20.6%, 15.1%, and 16% respectively.  So rather than
try to correlate electricity consumption to local energy regulations,
it is clear that the per capita consumption numbers by state are a much
better indicator of the presence of heavy industry. In other
words, the graph Drum shows is actually a better illustration of the
success of CA not in necessarily becoming more efficient, but in
exporting its pollution to other states.  No one in their
right mind would even attempt to build a heavy industrial plant in CA
in the last 30 years.  The graph is driven much more by the growth of
industrial electricity use outside CA relative to CA.

Now take the residential numbers.  Lets look again at the states at
the top of the per capita list:  Alabama, South Carolina, Louisiana,
Tennessee, Arkansas, Mississippi, Texas.  Can anyone tell me what these
states have in common?  They are hot and humid.  Yes, California has
its hot spots, but it has its mild spots too  (also, California hot
spots are dry, so they can use more energy efficient evaporative
cooling, something that does not work in the deep south).  These
southern states are hot all over in the summer.  So its
reasonable to assume that maybe, just maybe, some of these hot states
have higher residential per capita consumption because of air
conditioning load?  In fact, if one recast this list as
residential use per capita, you would see a direct correlation to
summer air conditioning loads.   This table of cooling degree days weighted for population location is a really good proxy for how much air conditioning is needed by state.  (Explanation of cooling degree days).
You can see that states like Alabama and Texas have two to four times
the number of cooling degree days than California, which should
directly correlate to about that much more per capita air conditioning
(and thus electricity) use....

OK, now I have saved the most obvious fisking for last.  Because
even when you correct for these numbers, California is pretty efficient
vs. the average on electricity consumption.  Drum attributes this,
without evidence, to government action.  The NY Times basically does
the same, positing in effect that CA has more energy laws than any
other state and it has the lowest consumption so therefore they must be
correlated.  But of course, correlation is not equal to causation.
Could there be another effect out there?

Well, here are the eight states in the data set above that the
California CEC shows as having the lowest per capita electricity use:
CA, RI, NY, HI, NH, AK, VT, MA.  All right, now here are the eight
states from the same data set that have the highest electricity prices:  CA, RI, NY, HI, NH, AK, VT, MA.  Woah!  It's the exact same eight states!  The 8 states with the highest prices are the eight states with the lowest per capita consumption.
Unbelievable.  No way that could have an effect, huh?  It must be all
those green building codes in CA.  I suspect Drum is sort of right,
just not in the way he means.  Stupid regulation in each state drives
up prices, which in turn provides incentives for lower demand.  It
achieves the goal, I guess, but very inefficiently.  A straight tax
would be much more efficient.

Via Hit and Run, this is an incredible site for stat-geeks to fool around.  Top 101 city lists.

#1 Average Sunshine!  I have also lived in the 4th least sunny city.  Sunnier is better.   Seattle is not among the rainiest in terms of total inches, because it never rains very hard.  If you could measure rainy as "number of hours per month that rain is falling", Seattle would be right up there.  In places like Houston, you get a lot more volume of rain, but you get a whole years worth in just a couple of hours.

Other interesting ones:

• Size of government:  Largest local government salary and wages per resident.  Every man, woman, and child in Cambridge MA spends $3,532 a year towards the salary of a City of Cambridge government employee.
• Property taxes per resident (look at the map, all but 8 of 101 in the northeast)
• Living your whole life in the same spot
• The average woman in St. Louis weighs over 165 pounds?
• Counties with a net exodus are still dominated by the upper midwest, but look at people moving out of coastal California!
• I have often argued that the numbers of people without health insurance are skewed by new and illegal immigrants.  Check out this map of places with the most uninsured - they all seem to border Mexico
• Chicago was named the windy city because of its politics, not because of actual wind

I just wish they had a better explanation of the metric and the data source for each