Archive for the ‘Political Science’ Category

Now that the 2016 election results are available by town for New Hampshire, I thought I would take a look at where libertarian candidates tended to do well or poorly, and how that pattern compared with conservative versus progressive support by town.

To measure libertarian voting by town, I used different variables in different years.

For 2008, I used:

  • Ron Paul vote share in the Republican presidential primary,
  • Libertarian Party vote share in the general presidential election (both Bob Barr and George Phillies had separate Libertarian candidacies in New Hampshire that year),
  • Libertarian Party vote share in the general gubernatorial election.

For 2012, I used:

  • Ron Paul vote share in the Republican presidential primary,
  • Libertarian Party and Ron Paul write-in vote share in the general presidential election,
  • Andrew Hemingway vote share in the Republican gubernatorial primary of 2014. (So this is sort of a 2012-4 measure really.)

For 2016, I used:

  • Rand Paul vote share in the Republican presidential primary (he had dropped out of the race, but over 1% of voters voted for him anyway),
  • Frank Edelblut vote share in the Republican gubernatorial primary,
  • Libertarian Party vote share in the general presidential election,
  • Libertarian Party and Aaron Day vote share in the general U.S. Senate election.

I also calculated conservatism vs. progressivism by town for 2012 and 2016.

For 2012 conservative voting, I used:

  • Republican Party general presidential election vote share,
  • Republican Party general gubernatorial election vote share (2012),
  • Republican Party general gubernatorial election vote share (2014).

For 2016 conservative voting, I used:

  • Republican Party general presidential election vote share,
  • Republican Party and Aaron Day vote share in the general U.S. Senate election,
  • Republican Party general gubernatorial election vote share.

I also looked at how these variables correlated with each other. The strongest correlations I found outside the Republican candidates’ correlations with each other were between LP + Day Senate 2016 vote share and Ron Paul 2012 primary vote share (r=0.51), LP + Day Senate 2016 vote share and Trump vote share (r=0.37), Rand 2016 and Ron 2012 primary vote share (r=0.32), LP + Day Senate 2016 vote share and Edelblut ’16 vote share (r=0.30), Johnson ’12 + Ron Paul write-ins and Ron Paul 2012 primary vote share (r=0.51), Trump vote share and Ron Paul 2012 primary vote share (r=0.45), Paul ’12 and Paul ’08 vote share (r=0.49), Paul ’08 and Hemingway ’14 vote share (r=0.43), and Paul ’12 and Hemingway ’14 vote share (r=0.34). Interestingly, the Johnson-Weld ticket really didn’t correlate with anything else at all, suggesting that most of their voters were simply anti-Clinton and anti-Trump, not libertarian leaners. Most of those voters probably won’t stick around for future Libertarian candidacies unless they don’t have a lot of options.

These correlations also imply that a lot of Ron Paul’s 2012 primary vote came from disaffected, non-conservative, potentially populist or nationalist Republicans. This is consistent with what I reported here on Pileus years ago about how Ron Paul in 2012 added a bunch of anti-establishment, moderate to liberal independents and Republicans to his libertarian base. It also suggests, perhaps, some reason for optimism about the Trump phenomenon. A lot of his voters are simply alienated and not that strongly ideological. If they could vote for both Ron Paul in 2012 and Donald Trump in 2016, these people must be ideologically flexible, even if consistently anti-establishment.

So where can you find libertarians, progressives, and conservatives in New Hampshire? These maps tell the tale. (Unincorporated townships and towns with fewer than 100 votes are not mapped.)

nhlib16 nhlib12 nhlib08 nhcon16 nhcon12

Please note that each variable is recentered each year, so that it is impossible to compare towns across years in any absolute sense (“this town is becoming more libertarian” is an inference you absolutely cannot make from these data), though you could make comparisons over time, relative to the average town in New Hampshire (“this town used to be quite a bit more libertarian than other towns, but now it’s only average” is something you could say based on these data). The recentering has to be done because different candidates run and are included in the calculations in different years.

Now then, where are the libertarians? The results aren’t greatly different from those I’ve reported before, with New Hampshire’s most libertarian towns generally lying in the Appalachian mountains and foothills of the western part of the state. However, in 2016 there is a notable change, with the libertarian center of gravity shifting southward to Cheshire and western Hillsborough counties. This surprised me a bit at first, but it makes sense once you consider that conservatarian Republican gubernatorial candidate Frank Edelblut is from western Hillsborough County and absolutely dominated the vote in his hometown and nearby towns, coming close to knocking off then-presumptive nominee Chris Sununu (now governor elect) statewide.

Republicans, meanwhile, are much stronger in the southeastern third of the state than elsewhere in 2012, but in 2016 they had more even support throughout the state, making clear inroads into the relatively deprived North Country. This is consistent with the general shift of the party to the nationalist radical right with accompanying loss of support among upscale groups and growth among the white working class.

Scatter plots make these phenomena clearer (click to expand).

The last scatter plot, in particular, shows that between 2012 and 2016 Republicans lost, relatively to the average town, in upscale towns like Hanover, Bedford, New Castle, Amherst, Hollis, and New London (most of these are strongly Republican towns, but Hanover is the most Democratic town in the state). Meanwhile, Republicans gained in the small towns of the North Country, places like Landaff, Berlin, Northumberland, and Millsfield. Libertarians (not necessarily Libertarian Party, but libertarianish candidates) look to have gained in western Hillsborough County, towns like Greenville, Mason, and New Ipswich (all heavily GOP). Again, this pattern really reflects the strength of Republican Frank Edelblut in his primary.

Read Full Post »

There’s been some debate about whether independent conservatarian candidate Aaron Day (former chairman of the Free State Project Board of Directors) cost Republican Kelly Ayotte her U.S. Senate seat at this past election. Skeptics point to the fact that Day and Libertarian Brian Chabot between them about equaled Gary Johnson’s vote percentage in the presidential race (Johnson-Weld got 4.2% in New Hampshire, Chabot and Day between them about 4.1%). They say Day merely siphoned off Libertarian voters from Chabot, not Republican voters from Ayotte.

They’re probably wrong. Here’s why.

First, let’s note that Clinton and Trump were the two least popular major-party presidential candidates in the history of polling. In particular, Trump was (and is) by far the least popular presidential candidate (and now, president-elect) that we’ve ever seen. Under those conditions, you expect an outsized third-party vote share in the presidential race. By contrast, Maggie Hassan and Kelly Ayotte, the two U.S. Senate candidates, were reasonably popular – and about equally so. So we should expect a small third-party vote share in the U.S. Senate race. The fact that the third-party vote share in the Senate race equalled that in the presidential race therefore suggests something else is going on – perhaps a particularly strong independent candidacy. Libertarian Chabot got 1.7% of the vote, while independent Day got 2.4% of the vote – so if one of them was the particularly strong candidate, it was Day.

To try to see whether Day was siphoning off Republican or Libertarian votes, I looked at town-by-town results for all 239 New Hampshire jurisdictions with voters in this race. I then looked at how the number of Libertarian and Republican presidential votes by town correlated with Day’s support. The results are in the figure below.

Aaron Day Vote Sources

The coefficient estimate on percentage of the vote for Trump-Pence in a town is 0.05 and is highly statistically significant, as you can see by the tiny confidence interval on the estimate. This estimate means that for every 20 additional Trump-Pence voters in a town, one additional voter cast a vote for Day in the Senate race.

The coefficient estimate on percentage of the vote for Johnson-Weld in a town is 0.08 but not statistically significant. This estimate implies that for every 12 additional Johnson-Weld voters in a town, one additional voter cast a vote for day in the Senate race — but again, we can’t be sure this is really any different from zero.

Day’s vote share was a little more than one-twentieth of the Trump-Pence ticket’s in New Hampshire. In other words, we can be reasonably confident that almost all of Day’s electoral support came from Republicans, not Libertarians.

To verify that Day specifically was the spoiler, not Brian Chabot, we can do the same exercise for Brian Chabot’s vote shares by town. The figure below shows those results.

Brian Chabot Vote Sources

Look at the difference! Chabot pulled overwhelmingly from Johnson-Weld voters. For every four additional Johnson-Weld voters in a town, Chabot got an additional one vote. Meanwhile, Chabot did siphon a few Republicans. For every 80 additional Trump-Pence votes in a town, Chabot got about one vote. Still, these results suggest that over half of Chabot’s support came from Libertarians, not Republicans.

Now, Kelly Ayotte lost by 1,000 votes, just over 0.1 percentage points. My estimates suggest that without Aaron Day in the race, Kelly Ayotte would have won her race by about 15,000 votes, similar to Republican Chris Sununu’s margin of victory in the governor’s race.

Is it possible that without Aaron Day in the race, those 16,000 or so disaffected Republicans would still have voted for some other third-party candidate or a write-in, or just that race blank? Sure, it’s possible. We can’t rule out that possibility because we can’t do an experiment in which we randomly assign some towns’ ballots to have Aaron Day on them and some not to have him on them. But at minimum these results strongly suggest that Kelly Ayotte disaffected a decisive share of Republican voters who went for Trump and Sununu but not for her.

Read Full Post »

I’ve started a new blogging gig at Learn Liberty, a project of the Institute for Humane Studies. I’ll be putting links to these posts here. My posts there will have the benefit of an editor, which is probably something I need.

The first is on partisan rationalization and why epistocracy may not save us after all – a suitable topic after this election, no? Here’s the lede:

Highly informed voters are also highly biased. That’s a serious problem for democracy, but also for any other system of political decision-making in big groups.

Two new books, Against Democracy by Georgetown moral philosopher Jason Brennan and Democracy for Realists: Why Elections Do Not Produce Responsive Government by Princeton University political scientist Christopher Achen and Vanderbilt University political scientist Larry Bartels, deal intellectual hammer blows to the political system so many of us take for granted: democracy.

Read Full Post »

On March 15, I had the opportunity to testify at the U.S. House Foreign Affairs Subcommittee on Europe, Eurasia, and Emerging Threats, chaired by California Rep. Dana Rohrabacher, on the topic of whether the U.S. government should change its policy toward national self-determination movements. I’m posting here my written testimony (my oral testimony had to be briefer). The conversations with the congressmen were interesting, and I think they revealed something about the foundations of U.S. policy toward such movements. Rep. Weber grilled me a bit about whether countries such as Spain, France, and Italy that prohibit secession in their constitutions should not enforce their constitutions. I responded that these countries should change their constitutions. However, another point I would make now is that the purpose of constitutions is to constrain governments, not citizens. Statutes and administrative regulations constrain citizens, but the constitution in turn restricts the ways in which government can create and implement these rules. So a constitutional prohibition against secession could really be enforced only against government agencies, and a government would be well within its legal rights to allow citizens to pursue secession, even in the face of a constitutional prohibition.


National self-determination movements seek greater self-government for a national minority, typically including the right to vote on forming a new independent state. Recent examples of successful self-determination movements include South Sudan, Kosovo, Montenegro, and East Timor. Ongoing self-determination claims are found in Scotland, Catalonia, the Faroe Islands, Kashmir, Tamil Eelam, Somaliland, Western Sahara, West Papua, Tibet, Mindanao, and many other places. Like other states, the U.S. government faces decisions about whether to recognize declarations of independence, to enter into diplomatic relations with new states, and to engage in diplomacy with other states about self-determination movements within their borders.

In my testimony, I will first describe the current state of self-determination movements around the world, then summarize what scholars have learned about the relationship between self-determination conflicts and violence. I will conclude by assessing the validity of claims advocating the creation of new states or changes to national borders.

The Current State of Self-Determination Movements

Self-determination movements generally take one of two forms: political parties and armed groups. Currently, secessionist political parties that seek at least a vote on independence are found in Belgium (Flanders), Canada (Quebec), Denmark (Faroe Islands and Greenland), Finland (Åland), France (Brittany and Corsica), Germany (Bavaria), Italy (Veneto and Sardinia), Spain (Catalonia, the Baleares, the Basque Country, Navarre, Canary Islands, and Galicia), the UK (Scotland and Wales), and the United States (Alaska and Puerto Rico). In addition, irredentist parties, which seek to move territory from one country to another, are present in the UK (Northern Ireland) and Italy (South Tyrol).

Armed self-determination movements are typically found in the developing world. Figure 1 shows where intrastate armed conflicts on territorial issues (generally, self-determination) occurred during the 2011-2014 period. These conflicts require at least 25 battle deaths in a single year to be counted.

territorial conflicts map

Developing countries usually forbid self-determination movements from organizing as political parties. For instance, Turkey, India, Pakistan, Sri Lanka, and Russia make advocacy for the self-determination of a particular region a criminal offense — an act that would be protected by the First Amendment in the United States.

Western, liberal democracies typically allow secessionist parties to organize and contest elections, but they do not all allow for secession. The Canadian Supreme Court has ruled that Quebec secession is negotiable if the province votes by “a clear majority on a clear question” for independence. The United Kingdom negotiated the terms of an independence referendum for Scotland and agreed to be bound by the result. The Danish government concedes a right to independence for Greenland and the Faroe Islands, and breakup is routinely discussed as a legal option for Belgium. St. Kitts and Nevis and Liechtenstein have constitutional clauses protecting the right of secession. On the other hand, France, Spain, and Italy all have constitutions explicitly defining their countries as indivisible, thus proscribing secession.

Majority support for independence in a population is rare. As of this writing, in all the high-income democracies of Europe, North America, and the Pacific Rim, there is only one region in which parties clearly favoring short-run independence have won an absolute majority of votes in any recent election: Scotland. Furthermore, in Scotland, many voters voted for the Scottish National Party (SNP) without favoring independence, and support for independence has been below 50 percent in polls since that election, including
the September 18, 2014 referendum itself. Using data from the Minorities at Risk project, I found that as of 2003, 107 ethnonational minorities, 38 percent of the total number in the data set, had a secessionist organization of any size (Sorens 2012, 56). In a recent article, I estimated the percentage of the population supporting independence in every state of India, finding figures no higher than 20 percent anywhere (Sorens 2014, 264).

The Causes of Self-Determination Conflicts

Popular demand for independence comes from a combination of a distinctive cultural identity, territorial coherence, and either political or economic benefits of independence (Sorens 2005; Hale 2008; Sorens 2012). Having just one of these elements is not enough, which is why the vast majority of minority nations around the world do not have any secessionist movement at all.

One worry about allowing secessionist movements is so-called “contagion” across regions or countries, but secessionism does not in fact seem to be contagious across countries, although it does have a tendency to spread within a country (Ayres and Saideman 2000; Sorens 2012), which is why governments often crack down on them (Walter 2006).

At the individual level, there is some evidence that voter support for independence is rational, that is, related in the expected way to the expected benefits of independence (Howe 1998). However, there is a difficult-to-resolve debate about the extent to which independence support is caused by voters’ assessments of the benefits of independence, or if instead independence support causes those estimates of benefits through a process of rationalization (Mendelsohn 2003).

Secessionism is strongly associated with violent conflict (Toft 2003). In general, separatist civil wars last longer than other kinds of wars, implying that the warring parties cannot find negotiated settlements even when the conflicts are stalemated (Fearon 2004; Sorens 2012).

I find that providing a legal path to independence is associated with less ethnonationalist rebellion (Sorens 2012). The United Kingdom, Canada, Denmark, and Belgium have had much less secessionist violence than France, Spain, and Italy — and secessionist violence has gone away in Puerto Rico since the U.S. government informally recognized their right to independence. Clauses permitting secession were also crucial to peace agreements ending the conflicts in Northern Ireland, South Sudan, and Bougainville (part of Papua New Guinea). The European Union’s Treaty of Lisbon explicitly recognizes member states’ right to secede from the Union, because no country would want to join a union they could never leave.

Implications for U.S. Policy

A legal path to independence can promote peace by constraining secessionists and central governments to pursue their aims through electoral and legislative means. On the one hand, secessionists have no excuse for resorting to violent tactics; to do so would be to admit failure to persuade a majority of the people they claim to represent, while imposing costs of violence on the very people they purport to represent and from whom they would have to recruit. On the other hand, central governments often cannot commit to respecting
a negotiated regional autonomy compromise without also conceding a right to secede. The South Sudanese and Bougainvillean secessionists would probably not have agreed to a peace deal without a referendum guarantee. These conflicts lasted 22 and nine years, respectively. Authoritarian and especially nationalistic central governments will face both desire and opportunity to renege
on previously negotiated autonomy arrangements; only a right to secede may be sufficient to deter them and thereby induce secessionist rebels to lay down arms in the first place. I also find that central governments permitting a legal path to independence are more likely to decentralize to ethnic minority regions and have never recentralized power in the post-World War II era (Sorens 2012).

If every country recognized its minority nations’ right to secede, only a few would apparently exercise such a right. Moreover, the overall level of global violence would likely decline by replacing intrastate conflicts with interstate conflicts. Intrastate conflicts are far more common than interstate conflicts (see Figure 2). Since World War 2, civil conflicts have killed seven times more people than interstate conflicts (Collier and Sambanis 2005; PRIO n.d.). Civil wars last much longer than interstate wars (Fearon 2004). Civil wars are also more likely to happen in more populous countries (Fearon and Laitin 2003). These findings suggest that a global increase in the number of independent states and a decrease in their average size would decrease the total number of conflict deaths.

conflict types

There are good reasons for the U.S. government to avoid assertively internationalizing other countries’ self-determination conflicts, which can look like meddling in other countries’ internal affairs. The U.S. arguably erred in refusing to negotiate a democratically authorized partition of Kosovo; as a result, an independent Kosovo lacks broad recognition from other states and is having trouble entering international institutions. Nevertheless, once a declaration of independence is issued, the U.S. government has no choice but to respond. In such an event, the U.S. government might wish to consider not only the interests of the host state, but also the interests of the seceding state and the effect of secession on regional stability. On average, replacing a state-to-nation relationship with a state-to-state relationship reduces violence.


Ayres, R. William & Stephen Saideman. 2000. “Is Separatism as Contagious as the Common Cold or as Cancer? Testing the International and DomesticDeterminants of Secessionism.” Nationalism and Ethnic Politics 6(3):92–114.
Collier, Paul & Nicholas Sambanis. 2005. Preface. In Understanding Civil War: Evidence and Analysis: Volume 1 (Africa), ed. Paul Collier & Nicholas Sambanis. Washington, D.C.: The World Bank.
Data on Armed Conflict. 2013. Peace Research Institute of Oslo. http://www.prio.no/Data/Armed-Conflict/, accessed September 20, 2013.
Fearon, James D. 2004. “Why Do Some Civil Wars Last So Much Longer than Others?” Journal of Peace Research 41(3):275–301.
Fearon, James D. & David D. Laitin. 2003. “Ethnicity, Insurgency, and Civil War.” American Political Science Review 97(1):75–90.
Hale, Henry E. 2008. The Foundations of Ethnic Politics: Separatism of States and Nations in Eurasia and the World. Cambridge, U.K.: Cambridge University Press.
Howe, Paul. 1998. “Rationality and Sovereignty Support in Quebec.” Canadian Journal of Political Science 31(1):31–59.
Mendelsohn, Matthew. 2003. “Rational Choice and Socio-Psychological Explanations for Opinion on Quebec Sovereignty.” Canadian Journal of Political Science 36(3):511–537.
Sorens, Jason. 2005. “The Cross-Sectional Determinants of Secessionism in Advanced Democracies.” Comparative Political Studies 38(3):304–326.
Sorens, Jason. 2012. Secessionism: Identity, Interest, and Strategy. Montreal, Que.: McGill-Queen’s University Press.
Sorens, Jason. 2014. “Legal Regimes for Secession: Applying Moral Theory and Empirical Findings.” Public Affairs Quarterly 28(3):259–288.
Toft, Monica Duffy. 2003. The Geography of Ethnic Violence: Identity, Interests, and the Indivisibility of Territory. Princeton, N.J.: Princeton University Press.
Walter, Barbara F. 2006. “Building Reputation: Why Governments Fight Some Separatists but Not Others.” American Journal of Political Science 50(2):313–330.

Read Full Post »

In the U.S., states have full authority over local government. Some states strictly centralize power and leave local government little to do. For instance, Hawaii has a single school district for the entire state, so that different localities cannot choose to spend different amounts on the government schools. Michigan effectively has a similar system, because it requires every school district to spend the same amount of money per student and redistributes tax funds across districts to make that possible. Vermont has also centralized school funding.

At the other end of the spectrum, states like New Hampshire let local governments pretty much decide their own level of funding for schools and other programs (about half of all local spending in the U.S. goes toward schools), and towns differ widely. If you want to live in a low-tax, low-spending town or a high-tax, high-spending town, it isn’t terribly difficult to find one. In the middle are states like Texas, where local governments are responsible for their own tax and spending decisions, but the most important level of local government is the county, much larger than the town, and it is therefore difficult to choose where to live based on local taxes and services.

Can we measure how decentralized each state is? I’ve tried to do so. The first measure of decentralization looks at how important local taxes are compared to state taxes. It divides local taxes by state and local taxes put together. This is a familiar variable to scholars of “fiscal federalism,” and it is typically called “tax decentralization.” Here is how the states rank on tax decentralization, as of fiscal year 2011-12, the most recent year for which data on local taxes are available from the U.S. Census Bureau:

New Hampshire 0.62475539
Alaska 0.584999114
Texas 0.555497037
Colorado 0.5420195
New York 0.540915308
Louisiana 0.520062304
South Dakota 0.514664958
Florida 0.508526077
New Jersey 0.503867865
Georgia 0.502739018
Missouri 0.490816162
Nebraska 0.486587041
Rhode Island 0.483462474
Ohio 0.47233672
Virginia 0.468452418
Illinois 0.466955731
Wyoming 0.465238453
South Carolina 0.459566438
Maryland 0.451067476
Pennsylvania 0.449406333
Arizona 0.440699694
Iowa 0.437082825
Oregon 0.434834984
Kansas 0.434319401
Washington 0.431347838
Wisconsin 0.423486277
Tennessee 0.421965652
Utah 0.420621904
Maine 0.411333699
Massachusetts 0.398031363
Connecticut 0.397670719
Montana 0.389680799
California 0.387518844
Nevada 0.383740954
Oklahoma 0.383081024
New Mexico 0.382245601
Alabama 0.382121115
North Carolina 0.366066432
Michigan 0.361458412
Indiana 0.352963108
Kentucky 0.33512693
Idaho 0.325219717
North Dakota 0.312465478
Mississippi 0.306727915
West Virginia 0.29895431
Minnesota 0.282530032
Hawaii 0.258739008
Arkansas 0.220173834
Delaware 0.215201394
Vermont 0.152464302

This isn’t the only way we can measure decentralization, though. After all, some states have more “competing jurisdictions” from which a prospective homeowner can choose than others do. To get at this concept was a little more complicated. I first counted the number of county, municipal, and township governments for each state from the U.S. Census Bureau. Then I looked at what proportion of local taxes came from each level of government and created a weighted average of number of local governments for each state. So if a state had 100 towns, 10 counties, 0 townships, and towns raised 20% of local taxes, while counties raised 80% of local taxes, the formula for the weighted average would be 10*0.8+100*0.2. The formula “rewards” states for letting lower-level, more numerous governments raise more taxes.

Then I thought about the decision of a homeowner in choosing a government to live under. Typically, your general location is set by where you have a job, say, a metropolitan area. But there may be several jurisdictions in that metro area to choose from. So I divided the “effective number of competing jurisdictions” described in the last paragraph by the state’s privately owned land area in square miles and multiplied by 100. So the resulting variable is the effective number of competing jurisdictions per 100 square miles of privately owned land. Higher values mean there is a lot of choice among governments.

Here is how the states come out on this variable measuring choice among governments:

New Jersey 5.619216533
Massachusetts 4.644661232
Pennsylvania 4.458726121
Rhode Island 4.016477858
Connecticut 3.634408602
Vermont 3.315789474
New York 2.934484963
New Hampshire 2.529344945
Wisconsin 2.189851779
Illinois 1.823655675
North Dakota 1.699505873
Delaware 1.586429725
Ohio 1.522032431
Maine 1.515194346
South Dakota 1.21988394
Missouri 1.105963152
Iowa 1.092652689
Indiana 0.972491305
Michigan 0.968708835
Kentucky 0.818674996
Minnesota 0.789141489
Arkansas 0.724807709
West Virginia 0.709066369
Oklahoma 0.693505752
Alabama 0.684705931
Georgia 0.551970462
North Carolina 0.535369811
Tennessee 0.506450581
Maryland 0.49052107
Kansas 0.479065166
Virginia 0.475682594
Florida 0.453352937
Nebraska 0.444783283
South Carolina 0.431428983
Louisiana 0.427531008
Utah 0.382243912
Mississippi 0.375744252
Washington 0.373979057
Texas 0.326573652
California 0.301273953
Colorado 0.284535146
Idaho 0.275746556
Oregon 0.273392409
Arizona 0.094351369
New Mexico 0.087975845
Montana 0.077669113
Hawaii 0.070909413
Wyoming 0.058851844
Alaska 0.042298043
Nevada 0.036335668

In general, the northeastern states score highly, largely because of a historical legacy of strong town government.

We can multiply both variables, tax decentralization and effective number of competing jurisdictions per 100 sq mi, together to get a single measure of how decentralized each state is.

New Jersey 2.831342639
Pennsylvania 2.003779755
Rhode Island 1.941816321
Massachusetts 1.848720839
New York 1.587307839
New Hampshire 1.580221888
Connecticut 1.44529788
Wisconsin 0.927372178
Illinois 0.851566468
Ohio 0.718911807
South Dakota 0.627831517
Maine 0.623250495
Missouri 0.54282459
North Dakota 0.531036915
Vermont 0.505539528
Iowa 0.477579724
Michigan 0.350147957
Indiana 0.343253553
Delaware 0.341401889
Georgia 0.277497088
Kentucky 0.274360038
Oklahoma 0.265668894
Alabama 0.261640594
Florida 0.23054179
Minnesota 0.22295617
Virginia 0.222834661
Louisiana 0.222342761
Maryland 0.221258101
Nebraska 0.216425781
Tennessee 0.21370475
West Virginia 0.211978447
Kansas 0.208067296
South Carolina 0.198270281
North Carolina 0.195980916
Texas 0.181410696
Washington 0.161315058
Utah 0.160780162
Arkansas 0.159583693
Colorado 0.154223598
Oregon 0.118880584
California 0.116749334
Mississippi 0.115251251
Idaho 0.089678217
Arizona 0.041580619
New Mexico 0.03362838
Montana 0.030266162
Wyoming 0.027380141
Alaska 0.024744317
Hawaii 0.018347031
Nevada 0.013943484

New Jersey is the state where the taxpayer has the most choice of government. While local property taxes are generally high there, that may simply reflect the preferences of local homeowners who want to spend money on services. It would be unsurprising if there are also some local jurisdictions in New Jersey where taxes are especially low.

In general, northeastern states, which are mostly left of center and high-tax, have a heretofore unseen advantage in their fiscal systems, letting competing local governments do much or even most of the taxation, making them responsive to local property owners. Perhaps it is precisely because of that responsiveness that overall tax burdens are allowed to be high in some of these states (New Hampshire aside): homeowner voters are more content with the way government uses their tax money there.

Read Full Post »

Updated to include two scatter plots

Having examined which states have the most and least libertarians, I’ve decided to do something similar for the 239 populated towns of New Hampshire. Towns are the most important level of local government here, and therefore the degree of libertarian-ness should make some difference to policy at the town level.

The indicators I use for number of libertarians are as follows: percentage of the vote for Gary Johnson and Ron Paul (write-ins) in the 2012 presidential general election (Ron Paul won a nontrivial number of write-ins in New Hampshire); percentage of the vote for libertarianish gubernatorial candidate Andrew Hemingway in the 2014 Republican primary (he got over 37% of the vote); percentage of the vote for Ron Paul in the 2012 Republican primary; percentage of the vote for Ron Paul in the 2008 Republican primary; and the percentage of voters registered “undeclared” (independent). These are all measured at the town level.

As in my research on the states, I use principal component analysis to reduce the correlations among these variables to a single “best” variable expressing their underlying commonality. I also “weight” the observations by population, since New Hampshire has many small towns, where sampling error should be higher (lots of zeroes and high percentages in election results). In fact, weighting the observations this way yields better results, as revealed by the eigenvalue of the first extracted component.

These variables do in fact correlate with each other and all contribute positively, as expected, to the extracted component. The highest scoring coefficient goes to 2012 Paul primary vote (0.55) and the lowest to undeclared registration percentage (0.25).

UPDATE: Here are two charts of Andrew Hemingway 2014 percentage against Ron Paul 2012 percentage, by town. The first limits to towns and cities with at least 700 population, the second to towns and cities with at least 10,000 population. As you can see, the correlation is strong.



And now for the lists of most and least libertarian towns…

Top 10:

Town Score
Richmond 11.2
Grafton 9.4
Wentworth 7.4
Alexandria 6.1
Lyman 6.0
Dorchester 5.7
Marlow 5.6
Clarksville 5.3
Croydon 5.2
Benton 5.1

Most of these are in Grafton County, where I also live. They are all small and rural. The most libertarian large town (over 5000 population) is Plymouth (score of 4.5), a left-leaning college town (also in Grafton Co.). The most libertarian-leaning municipality with a city form of government is Franklin in Merrimack County (score of 2.0). Almost all of the towns where libertarian candidates are most popular are in the west, especially northwest, of the state. Three exceptions are Francestown (5.0), Mason (4.3), Hill (4.0), and New Ipswich (3.9), but even these are west of I-93, which bisects most of the state. The top town east of I-93 is Pittsfield (3.2).

Here are the bottom 10:

Dixville -5.9
Hale's Location -4.7
New Castle -3.9
Rye -3.5
Jackson -3.2
Bedford -3.1
Waterville Valley -3.1
Atkinson -3.0
Stratham -3.0
New London -2.7

Four out of these 10 are in Rockingham County on the seacoast. Dixville and Hale’s Location are truly tiny. Bedford is a staunchly Republican suburb with a population over 20,000. In fact, many of the least libertarian places are well-to-do suburbs that are strongly establishment-Republican (Bedford, New London, Hooksett, Hampstead, Windham).

Examining the towns that are right in the middle of the spectrum will give us a sense of which places are most “representative” in their libertarian-ness. Here are those, filtering down to places with more than 1000 population:

Derry 0.2
Littleton 0.2
Goffstown 0.1
Keene 0.1
Manchester 0.1
Lee 0.0
Chester 0.0
Claremont -0.0
Sandown -0.2
New Boston -0.2

Some of these are not representative of the state in a left-right sense, however. New Boston, Goffstown, Littleton, and Chester are all firmly Republican, while Keene, Lee, and Claremont are if anything even more firmly Democratic. Derry (R-leaning), Manchester (D-leaning), and Sandown (R-leaning) could be considered somewhat representative of the state.

Read Full Post »

A few years ago, I did a statistical analysis of which states had the most libertarians, using data from 2004 and 2008 Libertarian Party vote shares and 2008 Ron Paul vote shares and contributions. David Boaz has prodded me to update these numbers in light of the 2012 election. This post does just that.

To come up with a single, valid indicator of how many libertarians are in each state, I use a technique called principal component analysis (PCA), which extracts the vector of data that best explains the correlations among multiple variables. Say I have a number of different measures of the number of libertarians by state. Using PCA, I can convert those different measures into a single measure. A crude way of doing this would be to simply standardize and average all of the different variables, but that method assumes that each variable is an equally reliable measure of the underlying concept. PCA actually tells us which variables are most reliable measures and weights them more heavily.

To see which states have the most libertarians, I use six measures: Libertarian Party presidential vote share in 2008 and 2012, Ron Paul contributions as a share of personal income in 2007-8, Ron Paul and Gary Johnson contributions as a share of income in 2011-12, and “adjusted” Ron Paul primary vote share in 2008 and 2012. Ron Paul vote shares are adjusted for primary vs. caucus, calendar, number of other candidates, and the like (for details see this post). Hawaii and Wyoming are excluded because they did not collect vote shares in the 2008 presidential primary. D.C. is included.

Here are the results of the PCA on these six variables:

. pca resid12 resid08 lp12 lp08 rpcpi08 libcpi12

Principal components/correlation Number of obs = 49
Number of comp. = 6
Trace = 6
Rotation: (unrotated = principal) Rho = 1.0000

Component | Eigenvalue Difference Proportion Cumulative
Comp1 | 2.81582 1.49201 0.4693 0.4693
Comp2 | 1.32382 .517957 0.2206 0.6899
Comp3 | .805859 .266932 0.1343 0.8242
Comp4 | .538928 .0754767 0.0898 0.9141
Comp5 | .463451 .411326 0.0772 0.9913
Comp6 | .0521252 . 0.0087 1.0000

Principal components (eigenvectors)

Variable | Comp1 Comp2 Comp3 Comp4 Comp5 Comp6 | Unexplained
resid12 | 0.1159 0.7527 0.1699 0.3288 0.5308 -0.0354 | 0
resid08 | 0.3400 0.5441 0.1240 -0.3297 -0.6750 0.0934 | 0
lp12 | 0.4360 -0.1868 0.3962 -0.6239 0.4133 -0.2408 | 0
lp08 | 0.3628 -0.3001 0.6360 0.5552 -0.1895 0.1724 | 0
rpcpi08 | 0.5218 -0.0665 -0.4366 0.2925 -0.1052 -0.6604 | 0
libcpi12 | 0.5263 -0.0897 -0.4513 -0.0152 0.2117 0.6828 | 0

“Resid*” is adjusted Ron Paul vote share, “lp*” is LP vote share, and the last two variables are contributions as a share of personal income. What this output tells us is that one single component has lots of explanatory power for the correlations among these six variables: we can interpret this component as the number of libertarians in a state. The method doesn’t give us a number interpretable as an absolute count of libertarians, but a number that we can interpret as representing how many libertarians each state has compared to all the others.

The second table of output shows how each variable contributes to each component. To the first extracted component, the one of interest to us here, the contributions variables actually contribute the most, while adjusted Ron Paul vote shares, especially in 2012, contribute the least. I have found elsewhere that in 2012 Paul did really well in states with lots of liberal voters, as he expanded his base beyond libertarians to antiestablishment liberals and moderates. As a result, his cross-state performance in 2012 isn’t actually a good measure of how libertarian each state is. Still, it contributes a little something to our measure.

Here is the extracted component, with all the states ranked from most to least libertarian:

state libertarians
Montana 5.504036
New Hampshire 4.163368
Alaska 3.586032
New Mexico 3.319092
Idaho 2.842685
Nevada 2.477748
Texas 1.632528
Washington 1.568113
Oregon 1.180586
Arizona 1.0411
North Dakota 0.7316829
Indiana 0.6056806
California 0.5187439
Vermont 0.4731389
Utah 0.2056809
Colorado 0.1532149
Kansas 0.107657
South Dakota 0.0328709
Maine -0.0850015
Pennsylvania -0.2063729
Iowa -0.3226413
Georgia -0.3296589
Virginia -0.3893113
Maryland -0.4288172
Rhode Island -0.470931
Tennessee -0.4882021
Missouri -0.4912609
Arkansas -0.5384682
Louisiana -0.5897537
Nebraska -0.6350928
Minnesota -0.7662109
Michigan -0.7671053
North Carolina -0.811959
South Carolina -0.8196676
Illinois -0.9103957
Ohio -0.9599612
Delaware -1.057948
Florida -1.072601
District of Columbia -1.091851
New York -1.225912
Kentucky -1.330388
Massachusetts -1.342607
Wisconsin -1.410286
New Jersey -1.431843
Connecticut -1.606663
Alabama -1.863769
Oklahoma -1.93511
West Virginia -2.244921
Mississippi -2.519249

Mississippi and West Virginia have the fewest libertarians, while Montana and New Hampshire have the most. Note that Montana and New Mexico will be overstated on this measure, because I have added half of the Montana Constitution Party’s vote share to the Libertarian Party vote share in 2008, because they listed Ron Paul on their general election ballot. No other state had the opportunity to run Ron Paul in the general election, however, so this choice overstates how many libertarian voters are in Montana. But excluding Ron Paul from Montana’s vote share would hurt them because he presumably drew lots of votes away from Bob Barr, the LP candidate, in that state. If I do exclude Ron Paul’s votes entirely from Montana 2008, then New Hampshire ends up just pipping them for most libertarian state. New Mexico is overstated because it is Gary Johnson’s home state, who did very well there both on contributions and on vote share.

These results are quite similar to those I found back in 2010, perhaps unsurprisingly since I included 2008 data on both occasions. Still, there are some small differences. New Hampshire has now easily passed Alaska for the #2 spot. Vermont, Maine, Kentucky, and Texas have gained, while Michigan, Idaho, Indiana, and Georgia have fallen.

Read Full Post »

Older Posts »

%d bloggers like this: