Feeds:
Posts
Comments

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.

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.

paulheming700

hemingpaul10k

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.

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.

New at e3ne.org, I discuss my conversations with high school students about the moral legitimacy of border restrictions:

We started our discussion with a little bit of improv theatre. I played a foreigner trying to get into the United States without documentation. Students volunteered to play a border guard trying to keep me out. Between us lay an invisible line, the border. I engaged them in a conversation about the moral justification of keeping me out.

To my surprise, the students were more confidently pro-immigration than I was! I played devil’s advocate some and tried to get them to appreciate the nuances of immigration policy.

My view is that borders are morally illegitimate because the state is morally illegitimate. Nevertheless, it can be permissible to use force to stop someone from settling in a particular area when doing so is necessary to safeguard public order or to preserve the minimal conditions for effective political autonomy for the existing communities in that area. For instance, I think it would be permissible for the U.S. government or an American state to prevent a large group of totalitarians from settling on their territory, provided the law does not provide a means for preventing them and their immediate descendants from obtaining citizenship. In a similar way, it would be appropriate for the Israeli government to prevent radical Arab nationalists from settling in their territory en masse. It’s also appropriate to exclude violent criminals, suspected terrorists, invading foreign armies, and, in the context of a welfare state, those unable or unwilling to work.

How can one group of human beings come to enjoy a right to enforce its authoritative commands on other human beings? In other words, how does government come to enjoy a right to rule, and how do citizens come to incur a duty to obey?

I consider the answer over at e3ne.org. The reasoning depends heavily on Michael Huemer’s book, The Problem of Political Authority, which I have reviewed here at Pileus. As a moral Lockean, my own view is that the U.S. government is illegitimate because it does not have a valid social contract with its citizens. That doesn’t mean the U.S. government is evil, or that we should try to overthrow it, but it does mean that the government doesn’t have any rights that ordinary citizens don’t also have. The U.S. government and its citizens are in a state of nature.

New at e3ne.org, I take up Peter Singer’s argument that we in affluent societies have far-reaching duties to aid the global poor, possibly to the extent of bringing ourselves down almost to their level. Excerpt:

Instead of buying a Starbucks coffee once a week, you could save that money – about $200 over the course of a year – and give it to a charity that saves lives. It’s morally wrong to buy Starbucks coffee when there are people dying around the world. Letting someone die so that you can enjoy Starbucks is like letting a child drown rather than getting your suit muddy.

It doesn’t matter that most other people aren’t living up to their moral obligations. Bystanders’ failure to save a drowning child doesn’t relieve you of a duty to save that child. If you can save a life without sacrificing anything morally significant, you must.

More here.

Recently I finished reading the book Gaming the Vote by William Poundstone. I also assigned part of it to my Ethics & Economics Challenge students. It’s a fun and informative read, draping heavy-duty political science in engaging story-telling. (My post at e3ne.org on the topic is here.)

The book’s central thesis is that the American electoral system is irrational, and that range and approval voting methods provide obviously superior alternatives to the plurality rule for “single-winner” elections. Along the way, Poundstone discusses the Arrow Theorem and why it provides no obstacle to making a comparative judgment among voting rules.

Arrow’s impossibility theorem says that no social choice rule (method for coming up with a social preference ordering over possible alternatives) can satisfy the criteria of non-dictatorship (no one person can make the decision for the whole group, irrespective of the preferences of the rest), universal domain (no preference orderings are simply ruled out of order), Pareto optimality (if everyone prefers X to Y, so should the group), and independence of irrelevant alternatives (changing your relative ranking of X and Z shouldn’t affect your choice between X and Y). In simple language, the Arrow theorem says that there’s no such thing as a “will of the people”: only individuals have preferences.

Poundstone takes issue with that interpretation of the theorem, arguing that the “independence of irrelevant alternatives” criterion should be relaxed or removed. In essence, Poundstone believes that we can make interpersonal comparisons of utility (utility as cardinal, not just ordinal), and that once we do so, we can come up with some social choice rules that are objectively superior to others, because they result in more aggregate human welfare.

The assumption of cardinal, interpersonally comparable utility lies behind the case for range (or score) and approval voting as alternatives to plurality rule. The former methods are said to result in less “Bayesian regret” when used either sincerely or strategically. For instance, plurality voting leads to the “spoiler effect” (Nader causing Gore to lose to Bush) and lots of tactical voting (Nader supporters voting Kerry instead). Sometimes it can even result in victory for a candidate that would lose by a majority to every other candidate, or simply fails to choose the candidate that would beat every other candidate (Poundstone discusses how Stephen Douglas likely would have won a pairwise majority vote in 1860 rather than Abraham Lincoln).

From page 239 in the Poundstone book comes this graphic based on plausible simulations of different elections under various voting rules:

Bayesian regrets under alternative voting rules

Bayesian regrets under alternative voting rules

Lower scores here are better, and thus you can see that range voting leads to overall least “regret” when voters are sincere. When voters are 100% strategic, range and approval tie (fully strategic range voters simply cast strategic approval-like votes: full marks to their favorite and the preferred candidate of the two with the best chance of winning, none to the rest).

Of course, the whole exercise depends on the notion that you can sum up regrets across voters. In some parts of life, we make rough-and-ready interpersonal comparisons of utility. When we speak of those “less fortunate,” we clearly have in mind the idea that the poor are less happy than the rich. The possibility of empathy seems to require a view that others are “more or less similar” to ourselves, including in their capacity for happiness. At the same time, the possibility of individuality seems to require that we acknowledge that others are “in some ways quite dissimilar” to ourselves. I can’t know what’s best for you, because your happiness has a large idiosyncratic, unmeasurable component.

Where does that leave us? The Bayesian regret calculations, it seems to me, give us good reason to favor range and approval voting over the current system, simply because in the absence of any other numeraire for making cost-benefit calculations of policies, the assumption of interpersonal comparisons of happiness, with everyone capable of the same amount of happiness (no utility monsters), is better than the alternative of throwing up our hands. But we still can’t get away from the fundamental insights drawn from the Arrow theorem: that only individuals have preferences and act on them, and that trying to maximize social welfare at the expense of respect for individuality is not only possibly unjust, but also irrational.

Follow

Get every new post delivered to your Inbox.

Join 1,057 other followers

%d bloggers like this: