29 November 2016

Topics Safe Enough To Risk Offending An Influential Academic Over

A new micro-symposium at SSRN provides counterpoints to and elaborations of University of Chicago law professor Richard Posner's essay, "What Is Obviously Wrong With The Federal Judiciary?" with only bylined pieces permitted.

The abstract stated:
After Richard Posner gave the Green Bag’s readers a double dose of “What Is Obviously Wrong With the Federal Judiciary,” we invited them to comment on his comments on the courts. So did he: “What I would most like to see would be criticism of the criticisms that I have made of the federal judiciary, and of the American legal culture more broadly, in this two-part article and at much greater length in Divergent Paths (and earlier books and articles). 
The Bluebook must have its defenders – let them defend their precious tome from me. And so must the awful legal jargon found in so many judicial opinions, and their verbosity; the superfluous headings and subheadings; the silly flourishes; the paeans to the adversary system; the pattern jury instructions; the standards of review; the dread of the italicized period; the spittoons behind the Supreme Court’s bench. The list goes on and on... But no; it seems I am to remain a voice crying in the wilderness. Pretty depressing.” This micro-symposium should inspire in Posner both a new cheeriness and more of the same old depression.
On which issues did academics (and perhaps a few practicing lawyers) dare challenge the judge?

* Which is the correct spelling of the latin legal phrase, "Ejusdem generis" or "Euisdem generis"?

* Headings and subheading in judicial opinions are important contrary to Posner's claim.

* An originalist method of interpretation of the constitution has some impact on the outcome of the case apart from the intended result of the person using it. Posner joins with an essay rebutting this and claiming that the so called originalist method advanced in the essay is "Faux Originalism".

My Two Cents On Other Matters

In Defense Of The Bluebook and the Italicized Period.

I was a Manual Of International Legal Citations editor during my term on the Michigan Journal of International Law, and in that post I gained some appreciation of the Bluebook (the leading legal citation style guide published by a consortium of three leading law school law reviews). 

Posner's University of Chicago style guide is its chief competitor in the marketplace for legal citation style guides, so he can fairly be accused of partiality.

What is good about the Bluebook?

When you write a law review or legal opinions or a legal brief, you must make stylistic choices about citation formatting whether you want to or not.  The decision to include a period in or outside italics, which is often hard to discern in print, is one such choice.

A good law review, court or law firm does so consistently and in a manner that has some outside justification for the choices made just like every other argument in persuasive legal writing. An authoritative ruling on every possible choice that could be made in formatting a citation, eliminates the need for the rest of the sane population of the world to debate these issues that must be decided to produce a finished work, but add little value of their own.

The core rules of the Bluebook are simple to apply in the handful of cases that make up 95% of legal citations, follow a common logic, and recognize the distinction between the more elaborate publishing register of an academic journal, and the more practical publishing register of a legal opinion or brief. The rest of the Bluebook provides guidance for myriad rarely used sources with examples that an author might otherwise fear to cite at all for fear that there is no proper legal citation for them. Thus, it implicitly broadens the scope of what is considered legitimate legal authority.

Also, the Bluebook's core rules (which I more or less follow in citations at this blog and its sister blog) thankfully eliminate much of the extraneous information required by ordinary academic journal style guides which are superfluous in legal writing, because, for example, the publishers of the leading case reporters are well known and the authorship of binding legal precedent is often irrelevant. 

The list of standard abbreviation conventions in the Bluebook is also particularly useful, and its jurisdiction by jurisdiction summaries provide a quick outline of the entire range of authoritative legal sources in existence.

In Defense of Legal Jargon In Court Opinions

Today, more than ever, legal researchers engage directly with source materials such as court opinions and statues, using word searches. 

In a word search, you only turn up cases that use the buzzwords of legal jargon you mention. An opinion that substitutes "plain English" for legal jargon will often be missed causing researchers to fail to identify relevant law since they no longer usually search using "digests" that organize case headnotes by subject matter regardless of the terminology used.

So, legal jargon in court opinions is more important now than ever before.

In Defense of Pattern Jury Instructions

At the end of every trial, the judge must devise jury instructions with input from the lawyers for both sides of the case relying on the relevant case law. There is infinite variety in how this can be done from scratch, which leads to constant legal battles over jury instruction language by counsel for both sides that is often critical on appeal because this is how all conclusions of law are implemented in a jury trial.

Yet, the vast majority of jury trials involve a modest subset of the entire law presenting legal issues that sometimes recur in every case, and sometimes recur in every case of a particular type.

Pattern jury instructions allow judges and counsel to dispense with wasteful argument over the details of jury instructions on matters where the law is settled that are tangential to the issues at hand for the jury (such as a lengthy description of their power to evaluate the evidence and the burdens of proof that apply). This frees up precious judicial and counsel resources at a time when a jury has heard all of the evidence and is waiting around to receive the case, to focus on the legal issues that are disputed and particular to the case at hand. It also insures relative uniformity across the judicial system.

In Defense of The Spittoons Behind the Supreme Court’s Bench

Why exactly would it be better for Supreme Court justices to spit on the floor than to spit into a spittoon? Duh!

And, if the spittoons were placed in front of the Supreme Court's bench, counsel presenting their cases to the court and the audience hearing oral arguments would be in line of fire of judicial spit. It is hard enough for them to face sharp questioning without having to dodge judicial spitballs as well.

In Defense Of Stating Standards of Review

Standards of review are obviously important to securing the right result in a case, and calling attention to this in each decision and brief, while somewhat redundant, does focus the authors on the task at hand in a manner that is often outcome determinative. Also, while the standard of review is clear to the participants even if not stated, it is helpful to state this so that clients and the public can better understand opinions.

In Defense of Verbosity

Verbose judges do so because (1) they seek to explain every element of their reasoning, and (2) it is faster to write a long opinion than a carefully honed smaller one. Both are desirable in the conditions they face relative to the common prior practice of dismissing arguments raised by counsel out of hand without any reasoned justification whatsoever.

Why Is Climate Change A Liberal Issue?

The need to address man caused climate change and the rising sea levels (among other things) that would result is largely a liberal issue in the U.S.  U.S. conservatives tend to be climate change deniers.

Part of this is historical conservative distrust of science and distaste for the government regulation of business necessary to protect the environment and slow climate change.

But, there is another piece to it as well. The impact of climate change due to rising sea levels which is its most obvious consequence falls disproportionately on the "coastal elites" of the United States who are predominantly liberal and vote Democratic.

Almost all of the major cities on the Pacific coast are at risk, with comparatively conservative San Diego fairing better than most Pacific coastal cities. The major cities of the Northeast, and Florida, which is one of the most liberal Southern states, also face grave flooding risks, as does New Orleans, a liberal hub within the South.

In contrast, rising sea levels disproportionately have little impact on "Red State America". Appalachia, the Midwest, Utah, Idaho, Arizona, the Great Plains and the interior of the South are far less affected. 

Of course, the correspondence is not perfect. The conservative Panhandle of Florida and the coasts of the Carolinas would be impacted, as would a lot of the Gulf Coast of Texas. Liberal outposts like Denver, Minneapolis, Chicago and Austin would be largely spared. But, these are the exceptions that prove the rule.

28 November 2016

Great Name For A Rock Band

Back in 865 CE, there was a Viking military group operating in Britain known as the Great Heathen Army according to contemporaneous chronicles, which would be a great name for a rock group.

25 November 2016

Wells Fargo Bank Still Violently Anti-Consumer And Crooked

Consumer arbitration is always a scam, but this one is particularly noteworthy, because the company claims the consumers are bound by agreements signed with signatures forged by the company. 

If there is any way that you can avoid doing business with this dishonest bank, you should take advantage of it and move your accounts immediately.
Even though disgraced Wells Fargo CEO John Stumpf has left the building, his most outrageous legal theories live on: on Wednesday, the company filed a motion in a federal court in Utah seeking dismissal of a class action suit by the customers it defrauded -- the bank argues that since customers sign a binding arbitration "agreement" when they open new accounts, that the customers whose signatures were forged on fraudulent new accounts should be subject to this agreement and denied a day in court.

This is the same argument that Stumpf made during his disastrous performance in front of a blazing Elizabeth Warren -- that Wells's poor customers should be subjected to agreements they never made, because Wells stole their identities and "agreed" on their behalf.
From here.

23 November 2016

More On Urban v. Rural Turnout In 2016

A county level data set shows urban America having about the same levels of support for Clinton and Trump respectively in turnout and partisan affiliation as for Obama and Romney, but rural and small town America shifted dramatically in favor of Trump and away from Clinton, relative to Romney and Obama respectively. 

There are also unconfirmed hints that a lot of this shift comes from discouraged rural Democrats and higher turnout among lapsed rural Republicans, rather than primarily from unaffiliated voters shifting their allegiances. 
Thanks to Rob‘s tip I was able to download all county level returns. I sorted out the counties in Census designated Urbanized Areas with at least a quarter of a million population (2010). This allowed me to look at the results for big cities (Metro America) and Rural America (rural +smaller metros). 
1) Metro American actually voted slightly more Democratic than in 2012 both in terms of the share of the vote and the total number of Democratic votes cast. In 2012 Democrats won Metro America 55%-44% and in 2016 they won it 54%-41%. The Democratic margin increased from 11% to 13%. Republicans won almost exactly the same number of votes but their share fell as Democrats added 2 million more voters. I think that the speculation about disenchanted Democrats not turning out is looking unsubstantiated based on these aggregate numbers. It also looks like most of those Republicans who said they weren’t voting for Trump came back to the GOP. It was almost a carbon copy of the 2012 election here. 
2) In Rural America Republicans increased both their vote share and their votes cast. Republicans gained 2.3 million more votes in Rural America and Democrats lost 1.9 million votes there. In 2012 Republicans carried Rural America 56%-42% but in 2016 they won it by a landslide margin 59%-36%. Democrats lost Rural America in 2012 but they didn’t get crushed there and that was the difference. Democrats lost ~2 million and Republicans gained more than 2 million. Where these Obama voters who switched? Or was this a case of 2 million discouraged Democrats staying home and 2.3 lapsed voters showing up and voting Republican? We really need someone to go through a rural county voter file to get answers. YouGov suggested it was 85% lapsed voters but I regard this as something to be confirmed. 
3) Bottom line is that if you lived in Metro America 2016 shaped up almost identical to 2012. What surprises me most about that is that Romney was a much more qualified candidate than Trump and yet Trump matched Romney’s performance. There was essentially no penalty in Metro America for breaking norms and other things that voters supposedly cared about. When we look at Rural America we see a dramatic shift from 4 years earlier. Honestly anyone who fails to focus on the shift in rural and small metro America isn’t looking carefully at the data. It’s very clear where the electoral shift took place.
Regional Breakdowns: Metro America.
2012 Dem 61% Rep 36% 2016 Dem 59% Rep 38% Northeast Metro.
2012 Dem 55% Rep 43% 2016 Dem 52% Rep 42% Midwest Metro.
2012 Dem 49% Rep 50% 2016 Dem 49% Rep 47% South Metro.
2012 Dem 57% Rep 41% 2016 Dem 58% Rep 35% West Metro.
2012 Dem 55% Rep 44% 2016 Dem 54% Rep 41% USA Metro.

Note: In the South Metro areas Democrats did better mostly because Republicans lost 3% to minor parties.

Regional Breakdowns: Rural America.
2012 Dem 50% Rep 48% 2016 Dem 43% Rep 52% Northeast Rural.
2012 Dem 42% Rep 56% 2016 Dem 32% Rep 62% Midwest Rural.
2012 Dem 38% Rep 61% 2016 Dem 34% Rep 63% South Rural.
2012 Dem 44% Rep 53% 2016 Dem 40% Rep 52% West Rural.
2012 Dem 42% Rep 56% 2016 Dem 36% Rep 59% USA Rural.

Note: Rural South used to be an outlier, but rural Midwest voted almost identical in 2016.
Matthew Gunning on Facebook via Steve Greene (emphasis mine).

I would also be very interesting in seeing the breakdown, within "Rural America" between small cities on one hand, and truly rural and small town voters. Other data from Florida suggest that Trump made more gains with suburban voters than with truly rural voters, which gives the story a very different feel and meaning.

Another data point is how late deciding voters broke:
The polls were wrong, and now Donald Trump is the president-elect of the United States. Everyone knows this. 
Except that's not the whole picture. Some of the polls were wrong to a degree, yes, but there was also something at work in the final days of the election: People who decided late broke strongly for Donald Trump in the states that mattered, according to exit polls. And without this apparent late surge, Hillary Clinton would be our president-elect — not Trump. 
In fact, if you look at the four closest states where Clinton lost — or, in the case of Michigan, where she's expected to lose — exit polls show late-deciding voters in each of them went strongly for Trump in the final days. In Florida and Pennsylvania, late-deciders favored Trump by 17 points. In Michigan, they went for Trump by 11 points. In Wisconsin, they broke for Trump by a whopping 29 points, 59-30.

And these weren't small groups of voters. The number of undecided and third-party-supporting voters who were still free agents in the final week was as many as 1 in 8 voters nationally -- an uncharacteristically high number for the eve of an election. (As Nate Silver noted, it was just 3 percent in 2012.)  
In Florida, 11 percent said they decided in the final week. In Pennsylvania, it was 15 percent. And in Michigan and Wisconsin — states where Trump made a late push — fully 20 percent of voters said they arrived at their choice in the last seven days. 
This isn't terribly surprising. As Philip Bump notes, we kind of saw it coming in the closing days of the campaign — particularly as backers of Libertarian nominee Gary Johnson migrated to Trump.  
What caused these voters to decide so late and break even more for Trump? Could it have been WikiLeaks or James Comey? (We're skeptical.) Or maybe Trump running a more disciplined campaign down the stretch? Or maybe it was just undecided voters breaking for the opposition party, as they are reputed to do. It's not clear why, but it seems they broke solidly in Trump's direction.

22 November 2016

Measuring Gerrymandering

The tricky part of gerrymandering is how to measure and analyze it.

How Do Partisan Votes Translate Into Partisan Seats?

One approach to measuring the impact of gerrymandering (arguably the most direct) is to compare voters cast for candidates of each party and compare them to the number of seats won and is called the Gallagher Index. (New Zealand provides a mix of single member district and proportional representation to mitigate gerrymandering incentives and keep the Gallagher Index low). Another similar index is the Sainte-Laguë method which is closely related to the Pearson's chi-squared test. One could use voter registration or previous detailed election results for some statewide partisan contest (e.g. the most recent elections for President or Governor) to measure new maps by this means in a hypothetical election conducted by this method.

Thus, when the electoral college chooses the candidate with a minority of the popular vote, particularly when the popular vote is not particularly close, there is gerrymandering as a result of how state lines are drawn by this measure. In the case of the electoral college, these disparities aren't common, but as the elections of 2000 and 2016 illustrate, they aren't particularly rare either.

The Efficiency Gap

A new measure relied upon in a new gerrymandering case in Wisconsin is the "efficiency gap": 
Courts have generally agreed that some partisan advantage in redistricting is tolerable, in part because voters themselves are not spread equally across a state or district by party. But the plaintiffs in the case, 12 state Democrats represented by the Campaign Legal Center, had argued that the Wisconsin remapping was among the most sharply partisan in the nation. 
Their lawsuit said that in the 2012 elections for the Assembly, Wisconsin Republicans won 48.6 percent of the two-party vote but took 61 percent of the Assembly’s 99 seats. A key question in Monday’s ruling, as in past challenges to redistricting, was whether that division was unacceptably partisan, a question that previous courts have stumbled over. 
“Nobody has come up with a standard to measure constitutionality — how to distinguish between malevolent, evil partisanship that’s manipulative, versus the natural advantage one party might have as a result of where voters happened to live,” said Edward Foley, the director of the Election Law Project at Ohio State University’s Moritz College of Law. 
In Monday’s ruling, the court was swayed by a new and simple mathematical formula to measure the extent of partisan gerrymandering, called the efficiency gap. The formula divides the difference between the two parties’ “wasted votes” — votes beyond those needed by a winning side, and votes cast by a losing side — by the total number of votes cast. When both parties waste the same number of votes, the result is zero — an ideal solution. But as a winning party wastes fewer and fewer votes than its opponent, its score rises. 
A truly efficient gerrymander spreads a winning party’s votes so evenly over districts that very few votes are wasted. A review of four decades of state redistricting plans concluded that any party with an efficiency gap of 7 percent or more was likely to keep its majority during the 10 years before new districts were drawn. 
In Wisconsin, experts testified, Republicans scored an efficiency gap rating of 11.69 percent to 13 percent in the first election after the maps were redrawn in 2011. 
Some experts said the efficiency gap gives gerrymandering opponents their most promising chance yet to persuade a majority of the Supreme Court to limit partisan redistricting. 
“It does almost exactly what Justice Kennedy said he was looking for back in the ’80s, a clear threshold for deciding what is acceptable,” said Barry C. Burden, the director of the Elections Research Center at the University of Wisconsin-Madison.
Per the New York Times (November 21, 2016).


The Washington Post's Wonkblog, meanwhile, focuses on a formula that maximizes the compactness of districts without regard to the political inclinations of the people in those districts, which it illustrates comparing the current national congressional district map to one drawn to optimize the compactness of the districts. 

But, it isn't clear how a compact district mandate would measure up in terms of gerrymandering impacts by other measures.

Interstate Partisan Competition

At the Congressional level, one theory that has been advocated in favor of not regulating gerrymandering at all, is that intense gerrymandering in blue states in favor of Democrats and intense gerrymandering in red states in favor of Republicans, on average, should balance out, maintaining a rough justice balance nationally without micromanaging the redistricting process in each individual state.

But, this approach works less well at the state level where state legislative districts are drawn by the same people who will be elected from those districts. There is no competition in that scenario.

Tight Kinship Networks and Corruption Go Hand In Hand

This study focuses on variations in corruption based upon cousin marriage rates in Italy and in international comparisons. Cousin marriage is widely believed to be one of the fundamental factors driving corruption in much of the Islamic world and South Asia, where cousin marriage is particularly common, in particular.

It is also notable in its own right that in parts of Southern Italy in the late 20th century, more than 50% of marriages were to a cousin.
Norms of nepotism and favoritism create corruption, subverting and disrupting impartial institutions and hampering economic development. However, the presence and strength of such norms varies widely within and between countries, and the literature has suggested that this variation is driven, in part, by ethnic fractionalization, with mixed results. 
We provide evidence for an overlooked -- but deep-rooted -- source of variation in corruption: sub-ethnic fractionalization, driven by mating patterns. The theory of kin selection provides a straightforward justification for norms of nepotism and favoritism among relatives; more subtly, it also implies that the returns to such norms may be influenced by mating practices. 
Specifically, in societies with high levels of sub-ethnic fractionalization, where endogamous (and consanguineous) mating within kin-group, clan and tribe increases the local relatedness of individuals, the relative returns to norms of nepotism and favoritism are high. In societies with exogamous marriage practices, the relative returns to norms of impartial cooperation with non-relatives and strangers are increased
Using cross-country and within-country regression analyses and a cross-country lab experiment, we provide evidence for this account. Our cross-country analyses show that corruption levels are robustly associated with consanguineous marriage rates, even when controlling for previously studied deep determinants of comparative development. Our within-country analysis exploits variation in consanguinity across Italian provinces and identifies the same pattern. Lab experiments in two countries with different mating patterns provide evidence for our proposed mechanism from subjects who interacted with kin, co-ethnics and strangers in a stylized corruption game.
Mahsa Akbai, et al., "Kinship, Fractionalization and Corruption" (October 3, 2016).

[G]eography- and culture-driven population division are the most common sources of sub-ethnic fractionalization. The influence of geography on population division should be obvious; as populations migrated around the world historically, they became isolated from one another due to vast distances and geographic barriers such as mountains, deserts and oceans that were only recently broken down by transportation technologies. Due to isolation, some groups accumulated relatively high local relatedness. 
However, our focus will be on culture-driven population division, in particular on the role of cultural preferences for and prohibitions of certain mating practices, which, when permitted, favor sub-ethnic fractionalization and thereby raise the relative returns to norms of favoritism and corruption. One mating practice that directly increases local relatedness and encourages sub-ethnic fractionalization is a preference for consanguineous marriage, most prominently among cousins.  In a cross-cultural ethnographic tabulation due to Murdock (1967) and Gray (1998), a total of 476 out of 1024 societies for which we have data either permitted or favored first and/or second cousin marriage, and estimates suggest that roughly 10% of marriages around the world today are consanguineous (Bittles, 2012). 
While the negative health effects of consanguinity are well-known (e.g. increased risk of autosomal-recessive disorders), some believe that there are countervailing positive effects as well (Jaber et al., 1998). In a small sample, genomic estimates of the “inbreeding coefficient”, which measures “the proportion of a genome that is ‘autozygous’ - homozygous for alleles inherited identically by descent from a common ancestor,” are correlated as expected with consanguineous marriage rates (r = 0.349, p-value = 0.04, N = 26). The correlation would likely be higher except that the genomic measures also capture background (so-called “cryptic”) inbreeding. 
The wide diversity in attitudes towards consanguinity in human societies partially originates in religious beliefs due to the common jurisdiction of religious institutions over marriage. . . . Of particular note are the fact that the Catholic Church has placed restrictions on cousin marriage since at least 500AD (sometimes extending these bans out to sixth cousins) and the fact that a persistent preference for cousin marriage can be seen in many Islamic countries, at rates as high as 50% of marriages. . . .
However, religion is not the only source of variation, and marriage norms may be particularly persistent, even as religious attitudes change. As one example, consanguineous marriage has long been prevalent in parts of Italy, despite it being an almost entirely Catholic country. Cavalli Sforza et al. (2004) suggest this may be a result of persistent cultural norms imported during the Arab conquest of southern Italy over 1000 years ago. Going the other direction, majority Protestant countries mostly legalized cousin marriage after centuries of living under the Catholic ban. Nevertheless, consanguinity remains rare in those countries. In the United States, cousin marriage is illegal in 25 states, though its frequency remains low even where it is legal. 
Others have argued that consanguinity is a cultural adaptation to social and ecological circumstances. In his history of the family, Goody (1983) suggests that consanguinity may be a property and wealth-preserving response to gender-egalitarian inheritance rules, which encourage the diffusion of property through out-marriage due to geographical population division (Pemberton and Rosenberg, 2014, p. 38). 
A few studies have examined the causes and consequences of consanguinity in small-scale societies using detailed genealogies to directly measure relatedness. Walker and Bailey (2014) show that among forager peoples, such marriages are rare due to norms of exogamy and fission/fusion dynamics that disperse kin across groups, but among agropastoralists, particularly those that practice polygyny, the practice is more common, with average spousal relatedness rising as high as r = 0.18 (almost 50% greater than first cousins, r = 0.125). Using log(surviving children) as a measure of fitness, estimates in Bailey et al. (2014) suggest that these marriage practices may be adaptive, with fitness maximized for moderate consanguinity among agropastoralists and with minimal consanguinity among foragers. Other evidence from Hoben et al. (2010) suggests that consanguinity may be more prevalent near the equator since it can raise the frequency of homozygosity for adaptive recessive mutations that defend against diseases and parasites, which are also more prevalent in warmer climes. 
Regardless of its diverse origins, wherever it is practiced, consanguineous marriage directly increases local relatedness and encourages fractionalization, thereby altering the returns to norms of favoritism and norms of impartial cooperation. Thus, variation in consanguinity rates facilitates a test of our main hypothesis: that sub-ethnic fractionalization causes corruption. To reiterate, our argument is that consanguinity increases incentives for kin altruism and hence for corruption, but the mechanism needn’t be genetic. That is, while it might be possible to show that differential selection pressures resulting from persistent inbreeding could increase the relative frequency of “altruistic alleles” favoring kin altruism, genetic change (or genetic difference) is not necessary to explain a change in the relative importance of kin altruism (and consequent corruption) across groups. Instead, we assume that all humans possess the capacity for kin altruism as well as for cooperation with non-kin, and that the norms operating in any given society merely favor one or the other mechanism of cooperation, to varying degrees. 
Since humans are social creatures reliant on cultural norms of cooperation and information-sharing for survival, selection pressures also operate on these norms, so that a society’s norms adapt to local conditions (Henrich, 2015). Thus, different mating patterns (due to culture or geography) may raise (or lower) the relative returns to norms of nepotism and local favoritism on the one hand and norms of impartiality, impersonal exchange and reciprocity on the other hand.
For instance, in countries Christianized for at least 500 years (and long-exposed to consanguinity bans), we see a significant decrease in the importance of clans and lineage groups (Greif, 2006). Moreover, cultural change and changes in local relatedness may be self-reinforcing. Historically, bans on consanguineous marriage were an important cause of migration, especially in agricultural societies where one male child inherited the family’s land. Siblings without property, especially females, had to migrate to find marriageable men (Cavalli-Sforza et al., 2004). This exchange of people across distances may have increased the returns to impartial norms facilitating peaceful interaction with strangers, perhaps in the long run encouraging the development of large-scale institutions based on generalized trust, such as markets, which also facilitate out-marriage. 
Inducements to consanguinity (or absence of a ban) may have opposite effects. By shrinking the sphere of social interaction to a group of more closely related people, consanguinity may encourage norms of partiality and in-group favoritism, and such norms may be readily applied to affines and other fictive kin, generating tight-knit local groups at the expense of potential impartial and impersonal arrangements that would expand the possibilities for exchange more broadly. Moreover, the presence of these relatively higher returns could have an amplifying effect, further raising the incentive to consanguineous marriage; see Jones (2016) on feedback between social norms and kin altruism. 
Our argument is related to the literature that distinguishes between generalized morality and limited morality (or amoral familism) (Banfield, 1958; Platteau, 2000; Tabellini et al., 2008). Limited morality is the extreme reliance on a narrow circle of family, friends or relatives; outside this circle, harming and cheating are allowed and frequent. In this narrow circle, people are raised to trust in-group members only. They are also taught to distrust people outside the circle, which hampers cooperation and exchange with strangers and outsiders, and as a result, impedes the development of formal institutions. Generalized morality is characterized by respect for abstract individuals and their rights, generalized trust and loyalty to general rules, which facilitates large-scale cooperation. The underlying mechanism that determines whether a society adopts limited morality has been attributed to strong versus weak family ties (Ermisch and Gambetta, 2010; Alesina and Giuliano, 2011, 2014), collectivism versus individualism (Yamagishi and Yamagishi, 1994; Yamagishi et al., 1998), and clan versus corporation (Greif and Tabellini, 2015; Greif, 2006). We also contribute to this literature highlighting another mechanism: sub-ethnic fractionalization increases the relative returns to limited morality and therefore encourages corruption.
According to their international regression model "a 1 standard deviation increase in consanguinity is associated with a reduction in quality of governance (i.e. increase in corruption) by about 0.75 standard deviations." Lower per capita income is also associated with corruption and explains almost all corruption effects otherwise attributable to geographical latitude.
A preference for consanguineous marriage is common in the Islamic world, and thus consanguinity rates and the percent of the country practicing Islam are highly correlated (Spearman’s ρ = 0.73, p-value < 0.001, N=67). In contrast, Catholicism has imposed a long-standing ban on consanguineous marriage, and this is evinced by a strong negative correlation between the share of a country practicing Catholicism and consanguinity (Spearman’s ρ = -0.57, p-value < 0.001, N=67). Finally, while Protestant religions do not officially ban consanguineous marriage, the frequency is quite low, and we find a large negative correlation between a country’s share of Protestants and consanguinity (Spearman’s ρ = -0.53, p-value < 0.001, N=67).
Meanwhile, political identity and group affiliations remain a dominant factor in American politics.

Does cousin marriage play a role in red state; blue state politics?

A list of states where cousin marriage is legal (per Wikipedia below) shows no clear red state, blue state pattern, but legality may not track the actual frequency of cousin marriage in the United States.

This is mostly because social factors keep the percentage of cousin marriages very low for the most part.

For the United States as a whole, the cousin marriage rate is about 0.2% (so there are about 250,000 people in these relationships), based mostly on old and not completely reliable data that are still unlikely to be wildly wrong, but "if you take a global perspective, consanguinity is not rare at all. Of the 70 countries studied, only 18 have consanguineous relationships as less than 1 percent of all marriages. In five countries, more than 50 percent of all marriages are between people who are second cousins or closer, and in Burkina Faso, it’s estimated that two of every three marriages are consanguineous."

Laws regarding first-cousin marriage in the United States
  First-cousin marriage is legal
  Allowed with requirements or exceptions
  Banned with exceptions1
  Statute bans first-cousin marriage1
  Criminal offence1

1Some states recognise marriages performed elsewhere, especially when the spouses were not residents of the state when married.clarification needed

Trump Wasn't Elected Because Americans Like Him

From here.

21 November 2016

Sometimes Weather In The Cloud Is Dismal

Technology related glitches in three distinct cloud based interfaces have each destroyed chunks of work that I have done in the last few days, one about 15 minutes worth due to a lost Internet connection midway through, one about half an hour's worth due to a lack of an "undo" option when an errant paste command erased a lot of other input, and one about three hours worth.

The two shorter losses could have been avoided if I had composed on a word processor (which has more safeguards) and then transferred the finished work online. But, doing that all the time would add more time than what I lose now and then to glitches. 

The longer loss (related to the state ICCES e-filing system) had a more complicated cause that was pretty much unavoidable under the circumstances. 

Three other glitches in a cloud based system (the various incarnations of the federal CMF/ECF e-filing system) have also caused no end of irritation and trouble over the past few months, basically due to the fact that the system required separate registration in the system for each federal court, requiring the use of inferior non-e-filing alternatives in two cases and a requirement for dual paper and electronic filing in other federal courts.

Still, it is always annoying in the extreme (to say the least), when it happens.

20 November 2016

Colorado Rocks Voter Turnout

Highest eligible voter turnout
1. Minnesota 74.2 percent
2. New Hampshire 72.6 percent
3. Colorado 71.3 percent 
Lowest eligible voter turnout
47(t). Tennessee, West Virginia 51.0%
49. Utah 46.4%
50. Hawaii 42.5%
From the Denver Post. The ultimate source of the turnout data is here.  Election results below are from CNN.

Nationally, eligible voter turnout was 58.4%.

Swing State Analysis

Trump won 306 electoral votes. He needed 270 to win, and 269 to tie sending the election to the U.S. House of Representatives where Trump would probably have won. Thus, if he had won 38 fewer electoral votes, he would have lost the Presidential race.

The marginal state that put Trump over the top was Pennsylvania where he won by a 1.2 percentage point margin.

What was turnout like in selected swing states and how did that compare to the results in those swing states?

Florida (Trump margin 119,770; 1.3 percentage points) 29 electoral votes

2016 Turnout 65.6% of VEP; disqualified 131,688 felons who were not currently incarcerated.

2012 Turnout 62.8% of VEP

2008 Turnout 66.1% of VEP


North Carolina (Trump margin 177,529; 3.8 percentage points) 15 electoral votes

2016 Turnout 64.9% of VEP; disqualified 5,747 felons who were not currently incarcerated

2012 Turnout 64.8% of VEP

2008 Turnout 65.5% of VEP


Pennsylvania (Trump margin 68,236; 1.2 percentage points) 20 electoral votes

2016 Turnout 61.4% of VEP; disqualified no felons who were not currently incarcerated.

2012 Turnout 59.5% of VEP

2008 Turnout 63.6% of VEP

Michigan (Trump margin 11,612; 0.3 percentage points) 16 electoral votes

2016 Turnout 65.6% of VEP; disqualified no felons who were not currently incarcerated.

2012 Turnout 64.7% of VEP

2008 Turnout 69.2% of VEP

Wisconsin (Trump margin 27,257; 1.0 percentage points) 10 electoral votes

2016 Turnout 68.3% of VEP; disqualified 66,222 felons who were not currently incarcerated

2012 Turnout 72.9% of VEP

2008 Turnout 72.4% of VEP

All turnout figures are persons voting for the highest office on the ballot. More details below the fold.

18 November 2016

How Common Are Very High Income African-Americans?

Klopfer, in an artful bit of statistical jujitsu manages to make some reliable inferences about high income African-Americans in the United States, which shows that the 1% is whiter than a moderate sophisticated intuition would expect. Full text of the paper does not appear to be available at this time, so the raw numbers which would be interesting in and of themselves, are not available.
I show that the Decennial Census and American Community Survey are valid sources of statistics on the top of the wage distribution, and replicate well-known findings from the administrative data commonly used to study income inequality. These surveys contain rich demographics that are not available in the administrative data, and I use them to provide the first race-specific estimates of top income earners in the United States. 
While the number of black Americans among the top 10, 5, and 1 percent of wage earners in the United States has grown sevenfold since the 1960s, the wages of blacks relative to whites haven't grown since the Civil Rights era. 
These facts are mechanically linked by rising wage inequality in the economy more broadly, and consequently the majority of progress by blacks into the top echelons of wage earners has occurred within industries, occupations, and levels of education, with comparatively little reallocation across these categories. 
Further, wages of blacks at the pth percentile of the black wage distribution, relative to those of whites at the pth percentile of the white wage distribution, are 70% from the 50th to the 95th percentile, but black relative wages fall to 50% by the 99th percentile of the respective wage distributions. I show that this relative wage slump occurs because top white wages follow the Pareto distribution and top black wages do not. Models that generate Pareto wages cannot explain the black distribution, but models of discrimination or human capital investment may.

Republicans Made State Office Gains In The 2016 Election

Results of the 2016 Election Continued

* Republicans lost two U.S. Senate seats (in the 100 member chamber) with one more to be decided later this year in Louisiana (and the gain of the Vice Presidential swing vote as well), but retained control (which is more meaningful now that the nuclear options has reduced the relevance of the filibuster). Republicans lost six U.S. House seats with four seats still too close to call, in the 435 voting member chamber, but again, retained control. 

These minor setbacks are overshadowed by the fact that Republican will control both houses of Congress, the Presidency and will soon secure a conservative majority on the U.S. Supreme Court. The fact that Trump won the Presidency by razor thin margins in three states for a non-landslide electoral college win (with a very close 16 electoral vote Michigan result still not final but leaning towards Trump) while losing the popular vote is basically irrelevant.

* At the state level, however, Republicans made modest gains. 

Republicans picked up three Governorships from Democrats brining the tally at 33-15 (with two sitting Governors who do not are not major party political officials). Republicans held 22 Governorships and Democrats held 28 immediately before President Obama was elected. Over eight years, Republicans have picked up 11 state governorships, while Democrats have lost 13.

Republicans picked up 46 state legislative seats from Democrats (out of a total of 7,370 seats nationally, including 49 in Nebraska's unicameral legislature where all state legislators are nominally non-partisan) bringing the tally to 4,170 Republicans, 3,129 Democrats and 22 independents and members of minor parties in the 49 states with partisan bicameral legislative chambers (with about a dozen state legislative races left that are still too close to call).
Republicans gained more than 700 [state legislative] seats in the 2010 midterm elections and nearly 300 in the 2014 midterms as Obama’s approval ratings suffered. Democrats clawed back more than 100 seats in 2012, when Obama won reelection. In total, Republicans control nearly 1,000 more legislative seats than they did when Obama took office. The Republican share of state legislative seats has grown from just under 44 percent in 2009 to 56 percent after Tuesday’s election. After the latest losses, Democrats will hold just 42 percent of [combined state and federal] legislative seats in the nation. . . . The GOP now controls the most legislative seats it has held since the founding of the party. 
Via The Hill.

They will now control 67 of the 98 partisan legislative chambers (they gained three and lost three this year), and also have de facto control over Nebraska's unicameral "non-partisan" legislative chamber, "after winning new majorities in the Kentucky House, the Iowa Senate and the Minnesota Senate. Democrats picked up control of both the state Assembly and Senate in Nevada, and the New Mexico state House. . . . Since Obama took office, Republicans have captured control of 27 state legislative chambers Democrats held after the 2008 elections."
Simply controlling more seats in a legislative chamber does not necessarily guarantee a majority. A coalition of Democrats, independents and a handful of Republicans will give Democrats control of the Alaska state House. Though they are in the minority in both states, Republican-led coalitions control the Washington and New York Senates.
Via The Hill.

Colorado Results

Colorado supported Hillary Clinton in the Presidential race (Colorado had been the marginal state for President Obama in both 2008 and 2012), and the Congressional delegation in Colorado (One Senator and four Representatives were Republicans, one Senator and three Representatives were Democrats) was unchanged in 2016. Democrat Morgan Carroll was defeated soundly by incumbent Mike Coffman in the nationally watched 6th Congressional District race (the only really competitive race in the state), just as he soundly defeated Andrew Romanoff in 2014, despite the fact that the district is seemly evenly split between Republicans and Democrats as voter registration data and this year's state school board race demonstrated.

In state races in Colorado in 2016, Democrats picked up 3 seats in the 65 member state house giving Democrats a 37-28 majority, up from the 34-31 majority they previously held (restoring them to the majority they held following the 2012 election), while Republicans held onto their 1 seat majority (18-17) in the state senate. (Colorado's Governor Hickenlooper is a term limited Democrat, so that race will be an open seat in 2018.) Democrats did pick up a majority on the state school board by flipping the 6th Congressional District seat on the board by a very slim margin.

Many states have a dominant political party. Colorado is not one of them. Since 1990, the Democratic tides in state elections have been mixed.  Years in which a single party had control of the legislative process in Colorado are highlighted in the summary below. In the 28 years from 1991-2018, Republicans will have controlled the process in four of those years and Democrats will have controlled the process in six of those years. In the other eighteen years, there has been divided control of the legislative process in Colorado.

Election 1990          State Senate  Republicans  State House Republicans (R control) Gov D
Election 1992          State Senate  Republicans  State House Republicans (R control) Gov D
Election 1994          State Senate  Republicans  State House Republicans (R control) Gov D
Election 1996          State Senate  Republicans  State House Republicans (R control) Gov  D
Election 1998          State Senate  Republicans  State House Republicans (R control) Gov R
Election 2000          State Senate  Democrats     State House Republicans (Split control) Gov R
Election 2002          State Senate  Republicans   State House Republicans (R control) Gov R
Election 2004          State Senate  Democrats   State House  Democrats (D control) Gov R
Election 2006          State Senate  Democrats   State House  Democrats (D control) Gov D
Election 2008          State Senate  Democrats   State House  Democrats (D control) Gov D
Election 2010          State Senate  20 D-15 R     State House  32 D-33 R  (Split control) Gov
Election 2012          State Senate  20 D-15 R     State House  37 D-28 R  (D control) Gov D
After 2013 Recalls  State Senate   18 D-17 R     State House  37 D-28 R  (D control) Gov D
Election 2014          State Senate  17 D-18 R     State House  34 D-31 R  (Split control) Gov D
Election 2016          State Senate  17 D-18 R     State House  37 D-32 R  (Split control) Gov D


Land And People Are Not The Same

First, keep in mind that the figures above do not make clear how many people are subject to Republican, Democratic, or mixed partisan control at the state level.


Second, some of these results are simply the result of an ongoing process of "realignment" as conservative Democrats have either switched to the Republican party or have been replaced by conservative Republicans, while moderate Republicans have been replaced by either Democrats or conservative Republicans. In other words, partisan labels have become more descriptive while the actual liberal-conservative shift in the ranks of the nation's elected officials have been less extreme than it appears superficially looking only at party labels.

But, the difference between conservative Democrats or moderate Republicans who used to be more common, and liberal to moderate Democrats and conservative Republicans, who have almost completely replaced legislators of all other party/ideology combinations, are not trivial or irrelevant.

The deeper ideological divisions between the two major political parties has made compromise much more difficult, particularly from the Republican side which has moved much more sharply to the right than the Democrats have moved to the left.

Conservative Democrats, historically, were fairly aligned with other Democrats on neo-New Deal type economic policies, while parting ways on issues of social issues, civil rights and national security. Moderate Republicans tended to be center-left on social issues and center-right on economic issues. There used to be a few federal legislators in each party who overlapped with some members of the other party on an overall liberal-conservative scale. Now there are none.

The lack of any ideological overlap, the hard right ideological shift of the Republican party, and much of the Republican party's abandonment of many traditions of political civility and comity, has left a scorched earth of the demilitarized zone between the two parties has made 21st century American politics decidedly more ugly than it has been for a long time.