Update and Further Analysis of the New Hampshire Legislature

With this post, I’m reporting updated results on the ideological ideal points of New Hampshire legislators, introduced previously here. In that analysis, I found that libertarians in the New Hampshire House in 2014 tended to vote with the right (and vice versa) on most roll-call votes scored by the New Hampshire Liberty Alliance. That included votes on bills to prevent localities from acquiring military vehicles for police, to reform civil asset forfeiture, to protect public school students’ privacy, and other civil-liberty issues where you would have expected the left to be more libertarian. In fact, the more left-wing you were, the more likely you were to oppose the libertarian position on those bills.

However, there were a few bills on which the left was more libertarian, mostly dealing with marijuana and other criminal-justice issues. When those were separated out, the analysis revealed a distinct cluster of 10-40 legislators (depending on the strictness of the criteria for inclusion) who tended to vote with the left or center on this just-mentioned subset of social issues but with the center or right, respectively, on the majority of roll-calls: a libertarian(-ish) caucus.

I am now updating the analysis with 2013 roll-calls included, to cover the entire biennial session. I’m also reporting more charts and tables for the geekily inclined.

To recap, I ran a Bayesian IRT analysis with imputation of missing data (abstentions and absences) using R package “pscl.” First, I began with the hypothesis that libertarianism-communitarianism was the first dimension underlying all roll-call votes. This hypothesis seemed to work for most roll-call votes, but it failed on some. In fact, a left-right dimension underlies most legislators’ voting decisions. So I separated out the bills on which the hypothesis failed and ran separate analyses on both sets. That resulted in two dimensions of ideology: how right-wing you are on right-libertarian issues (henceforth, “right-libertarianism”) and how left-wing you are on left-libertarian issues (henceforth, “left-libertarianism”).

I also corrected some errors in the NHLA data (votes coded the wrong way and one individual legislator vote miscoded) and made some different judgment calls from them. I mentioned some differences I had with their inclusion of votes against casino bills as pro-liberty and votes against a domestic violence bill as pro-liberty. I would have dropped them from the analysis if they had made any difference to the results, but they didn’t, so I didn’t bother. However, I did make some more substantial judgments. I dropped two voter ID/registration bills that the NHLA supported (loosening voting requirements). These were party-line votes: libertarian Democrats voted with their party in favor, and libertarian Republicans with their party against. NH’s voter registration rules are extremely lax, and the reform proposals so modest, that I could hardly count these bills, which undid some changes of the 2011-12 legislature, as clearly pro- or anti-liberty. Had I included them, they would have dominated the second dimension, reducing the additional information it supplies beyond mere left-right ideology. Finally, I counted votes in favor of a bill banning prison privatization as anti-liberty and votes against expanding the research and development tax credit as pro-liberty, directly contradicting the NHLA’s positions (but with good reason, I think – the data showed stronger fiscal conservatives voting against the NHLA on both).

There were 404 legislators that served at some point during the 2013-14 session (the maximum at any one time is 400). There were 149 total bills I looked at, but in the end 136 made it into the right-libertarianism analysis and 11 made it into the left-libertarianism analysis.

Another small difference from the last effort is that this time I ran a NOMINATE analysis first to get priors on each dimension for each legislator. The NOMINATE analysis also suggests 2 dimensions of ideology, just rotated slightly differently.

For the geekily inclined, here is a table of the discrimination parameters for the 10 most important bills contributing to the right-libertarianism dimension (here’s the intuition: a higher discrimination parameter means the bill is more important in distinguishing the right-libertarian from the left-communitarian):

bill	mean	stdev	lower	upper proliberty antiliberty notvoting
hb544	8.27	1.159	5.892	10.343	154	182	68
sb413	6.284	0.894	4.848	8.267	132	202	70
hb271	5.065	0.793	3.508	6.798	155	206	43
hb370	4.19	0.57	3.136	5.281	151	188	65
hb1570	4.048	0.536	2.894	5.226	142	161	101
hb1541	3.939	0.541	3.025	5.418	109	162	133
cacr1	3.938	0.523	3.123	5.086	149	206	49
hb1403	3.775	0.531	2.676	4.933	118	173	113
hb427	3.654	0.505	2.772	4.661	152	185	67
sb120	3.545	0.458	2.756	4.544	119	186	99

All of these were losses for liberty, and given the Democratic control of the House then, that’s not surprising. But only two of them passed both houses and were enacted into law: SB 413, Medicaid expansion, and SB 120, increasing campaign finance reporting and registration requirements. The top bill, HB 544, would have created a state-based Obamacare exchange.

Here are the bills that fed into the second dimension, left-libertarianism vs. right-communitarianism (all 11):

bill	mean	stdev	lower	upper proliberty antiliberty notvoting
hb573	10.171	0.24	9.735	10.804	286	64	54
hb492	2.092	0.297	1.531	2.94	170	162	72
hb1625	1.492	0.218	1.118	1.99	215	92	97
hb621	1.376	0.177	0.946	1.669	193	136	75
hb1325	0.872	0.147	0.56	1.162	66	219	119
hb249	0.807	0.149	0.566	1.105	266	68	70
hb1237	0.701	0.092	0.551	0.9	231	97	76
hb1577	0.674	0.124	0.475	0.943	209	116	79
hb1501	0.526	0.097	0.253	0.728	211	86	107
sb296	0.429	0.095	0.283	0.638	210	128	66
hb1624	0.327	0.109	0.147	0.567	256	40	108

HB573, medical marijuana legalization, dominates this list. The next on the list, HB492, was marijuana legalization. All of these were victories except Death with Dignity, HB 1325, again predictable given Democratic control of the House. Only medical marijuana and HB1624, modernizing the juvenile justice system, were passed and enacted into law. Of course, some of the victories consisted in defeating bad bills.

Here were the 10 most right-libertarian legislators in the last session:

name	party
Groen Warren	R
Sylvia Michael	R
Meaney Richard	R
Hoell J.R.	R
Murphy Keith	R
Sandblade Emily	R
Notter Jeanine	R
Baldasaro Alfred	R
Lambert George	R
O'Brien William	R

William O’Brien was the 2011-12 Speaker of the House. Several of these legislators are Free State Project movers.

Here were the 10 most left-libertarian legislators in the last session:

name	party
Winters Joel	D
O'Flaherty Tim	D
Vaillancourt Steve	R
Ketel Stephen	D
Levasseur Nickolas	D
Bickford David	R
Gardner Janice	D
Friedrich Carol	D
Arsenault Beth	D
Carroll Douglas	D

Joel Winters was the first FSP mover elected to the state house (2006). Steve Vaillancourt is a left-libertarian gadfly and served in the legislature for close to two decades, before losing this year.

I created an index of libertarianism weighted 3:1 toward right-libertarianism (on the theory that economic and personal freedom are equally important, and right-wingers are better on economics and half of personal freedom, and worse on the other half). Here are the most libertarian legislators according to that blend:

name	party
Lambert George	R
Sylvia Michael	R
Garcia Michael	D
Hoell J.R.	R
Sandblade Emily	R
Pratt Calvin	R
Warden Mark	R
Meaney Richard	R
Murphy Keith	R
Tasker Kyle	R

Here is a plot of the legislators on both dimensions, color-coded by party (click to expand):

Note that a handful of communitarians do now reveal themselves: William Butynski, Daniel Hansberry, Katherine Rogers, Leigh Webb, Deanna Rollo, Richard Eaton, Mary Nelson (all Dems).