In a recent series of posts (finale here), I estimated the size of the liberty constituency in each state by conducting a principal components analysis of four variables: per capita donors to the Ron Paul campaign, unexplained Ron Paul primary vote share, mean Libertarian Party presidential vote share 1996-2004, and Libertarian Party presidential vote share in 2008. Then, I regressed estimates of individual freedom from the Ruger-Sorens Index on the estimate of liberty constituency size, finding that states with larger liberty constituencies have more freedom.
Some conversations have raised the possibility of using these regression results to predict the effect of the Free State Project (see an exchange with Patri Friedman here). The idea is that I could plug in hypothetical values for New Hampshire on all the four variables above, assuming that they get 2,000, 5,000, or 10,000 more libertarian activists. In other words, given my estimate of the effect of the size of the liberty constituency on freedom, what would be expected to happen to freedom in NH if the liberty constituency in that state grew?
To do this, I first added 2000 Ron Paul donors and Libertarian Party voters to the state, along with 4000 Ron Paul voters. (Why? Because elsewhere I’ve found that each additional Free Stater in a New Hampshire town generated two additional votes for Ron Paul. Now, of course, there probably won’t be another Ron Paul campaign, but there will be other ways in which libertarian constituencies evince themselves – remember, I’m just trying to get a reliable measure of the size of the liberty bloc; the inputs as such don’t matter.) Then I figured out how that would change the estimate of the size of the libertarian bloc in New Hampshire. The aggressive assumption behind this move is worth noting. In particular, I’m assuming that the overall ideological distribution in NH on the libertarian-populist dimension shifts as a result of these activists. In other words, the overall relationship behind number of activists and size of liberty constituency remains constant – the liberty activists don’t just become an ideological ghetto. Presumably, it will take some time for activists moving into the state to have an effect on the mindset of the people already there, so the estimates I’m giving here are for a fairly optimistic, long-run-ish view of what the FSP can accomplish. (On the other hand, I’m assuming that the state remains fairly liberal on the left-right spectrum, which might be a pessimistic assumption.)
Now that I have a hypothetical value for New Hampshire’s future liberty constituency, I can plug that into the regression equation to see what value of freedom pops out, assuming that New Hampshire remains the same on every other variable. I do this by running 1000 simulations of the freedom regression, so that I can pull out a margin of error. When I do this, I find that New Hampshire’s expected freedom value increases by 0.37 on 0-1 scale, with a 95% confidence interval of 0.017-0.685. OK, what do those numbers mean? Well, that’s roughly the difference between New Hampshire and Nebraska today, or between Nebraska and New Jersey. That’s still pretty abstract, though.
To get a more concrete sense of what that means, I played around with the state policy data to see what changes would correspond to that kind of increase in freedom. I cut state and local spending by 3% of personal income (from 17.3%) and state and local taxes by 1.5% of personal income, from 8.4% (these don’t match up, because federal grants cover roughly half of state spending). Then I cut government employment by 2% of the workforce, from 10.8%. Then I gave New Hampshire Alaska’s gun laws (concealed carry without a permit and removal of some other minor regulations). Then I completely privatized wine and liquor and cut beer taxes to zero. Then I completely legalized marijuana possession, legalized cultivation and sale of medical marijuana, and decriminalized cultivation and sale for recreational use. (No state is actually this good.) Then I completely deregulated homeschooling: no testing, recordkeeping, or even notice required. (No state is actually this good.) Finally, I gave New Hampshire same-sex marriage, because, well, it already has that – but it got it after our data came out. That got New Hampshire up to the expected level of freedom after having had 2000 activists move in.
Now, there is a good bit of uncertainty about this estimate. It could be that these 2000 activists will have a much bigger or much smaller effect on freedom. The bottom end of the 95% confidence interval corresponds to just same-sex marriage, the gun law change, and medical marijuana. Not a huge deal. The upper end corresponds to everything mentioned, plus adding right-to-work, adopting the best existing occupational licensing regime in the country (Indiana’s, just 20% of the workforce licensed, compared to NH’s 23%), reducing victimless crime arrests by about 50%, to Hawaii’s levels, repealing all smoking bans on private property, abolishing cigarette taxes, legalizing prostitution, abolishing all campaign finance regulations, and cutting state and local debt burden by about a half. By that point, New Hampshire starts to look like a mix of Amsterdam and Alaska on personal freedoms and Hong Kong on economic freedom.
What about if 10,000 activists move to NH? Well, the freedom regression model doesn’t build in diminishing returns, so the simulations yield a predicted change in freedom of 1.46, roughly five times that predicted for 2,000 activists, unsurprisingly. At that point, we’re talking about cutting government to the bone, including tax and spending reductions of 50% or more and abolishing all of the remaining petty, paternalistic restrictions on freedom that we code, from gambling laws to sobriety checkpoints, legalizing assisted suicide, and completely deregulating education, complete with a strong tax credit-based school choice program, in addition to everything previously mentioned.
But at this point we are so far outside the range of observed politics that I strongly caution against taking these inferences all that seriously. It is possible to push a regression model much further than it can bear. I simply wish to get a sense of the orders of magnitude of change that might be possible with the FSP’s success.
In conclusion, it appears that with 2,000 activists who are smart about educating the general voting public and do not isolate themselves, the long-term gains to freedom in New Hampshire could be fairly extensive, though not approaching what anyone would describe as a “libertarian utopia.” It’s at about 4,000 effective libertarian activists that we could reasonably expect NH to start to look like the Amsterdam/Alaska/Hong Kong hybrid.