The Correlation of a State’s Political Affiliation on Various Metrics

At the beginning of every year, for 38 years now, United Van Lines has published its Annual National Movers Study. This study shows the aggregate of all interstate moves that have been done with United Van Lines moving trucks. The interpretation of this data can be taken a lot of ways. For 2014, the “big losers,” or those states where outbound moves highly outnumbers inbound moves, were New Jersey, New York, and Illinois. This study has been examined a lot in the press, with this article from Forbes being the one that put me onto the UVL report.

The Forbes article, specifically, cites the emigration factors being high housing costs, cold weather, and greater than normal impacts from the recession. The first factor, high housing prices, I can agree with, but the second two factors sound like some BS to me. First, there are a lot of cold states. In fact, in order of coldest to hottest states, New Jersey (27^th), New York (14^th) and Illinois (29^th) could be called “middle-of-the-road” at best. Plus, that just sounds ridiculous (“Honey! Call the realtor—it’s too damn cold here!”) Also, regarding the recession being cited as a factor, well, I’m not sure how valid that is. Yale University did a study ranking states 1-50 showing the impact of the recession (1 being least impacted) on citizen’s economic security (the ESI index) based on reduction in household income. New Jersey was 15^th, Illinois 24^th, and New York 33^rd.  So, in my opinion, blaming the recession for folks leaving the state six years later is pretty much just lazy reporting.

So, then, what did cause it?

The 2014 elections are just now in the rear-view mirror. The State of the Union is but two weeks away and, I must say, politics is on my mind. When I saw these three states (Illinois, New York, and New Jersey) as the big losers in migration, I couldn’t help but note that these three could be considered three overwhelmingly (if not the most overwhelmingly) democratic states in the U.S. Now, I’m not a democrat or a republican, and I’ve argued that being open-minded and tolerant in every debate is undoubtedly the only reasonable political position to have. But, I thought it’d be a good thought experiment and a test in my statistical savvy to see the correlation of a state’s political affiliations with various other factors—economic, migration and otherwise.

To judge political affiliation, I’m using Gallup’s “Conservative Advantage” number. This is simply the gap between the percentage of conservatives in a state versus the percentage of liberals (the farther right a point is, the more conservative the state). Alabama is the most conservative (36%) while Massachusetts is least conservative (-2%). I left out DC (-20%) because good data just doesn’t exist for most of what I was testing. I adjusted the Conservative Advantage to get only positive values so that the charts would be across just one quadrant, but this changes nothing regarding correlation or trends—it’s just a graph-shift from left to right.

To parameters I wanted to look at were as follows (with sources)

• Interstate Migration (using the same United Van Lines study)
• Fiscal Conditions (Mercatus.org)
• Taxes (WalletHub)
• Cost of Living (BLS, compiled by top50states.com)
• Corruption (State Integrity Investigation)
• Happiness (Time Magazine)

I’ll offer my prediction I had before I did any analysis, then the correlation, then a brief explanation:

Interstate Migration

Prediction: After seeing the headline states of who were the big losers in the migration battle, my assumption was that this was probably a trend—even if just a weak one—that big democratic strongholds probably saw some meaningful and negative population movement last year.

Results: Wrong! No Statistical Correlation. That point way up high is Oregon, where over two-thirds of all moves were into the state. Those three at the bottom are those three states that we talked about earlier (IL, NJ, NY).

But, apart from those three states, it looks like the scatterplot is pretty scattershot (see what I did there?). That R-squared number (.0055) indicates the strength of the correlation—a 0.55% correlation is essentially nil. My prediction was wrong, people aren’t really moving states based on political affiliation—that small sample size being all democratic states in the headline was just an anomaly and coincidental (and it led me to a faulty assumption).

Fiscal Conditions

Prediction: It’s no secret, democratic governments tend to want to spend more, have more lucrative public sector pension systems, larger social safety nets, and more “investment programs”. We can discuss the motivations about these issues until the cows come home, but I’d rather not. I predicted that the average fiscal condition (defined by Mercatus in a 1-50 composite rank as the ability to pay for personnel and services as well as debt levels and credit rating) in more democratic areas will probably be worse than in republican areas.

Results: Correct! With a strong statistical correlation. This is probably the only time my prediction will be right, but, wow, look at it:

The higher up you go, the weaker a state’s fiscal condition is. You can’t really argue with the correlation—it’s there—the “conservativeness” of a state explains half of its fiscal condition. And, states that are “more republican” tend to have stronger fiscal conditions.

Taxes

Prediction: This is in the same vein as the previous. I would certainly expect more democratic states to have a greater tax burden on their citizens (as defined by WalletHub as average state and local taxes paid per year). I’m not saying this is a good thing or a bad thing (although I’d tend to believe it to be a bad thing). But, just as I would predict more conservative states probably have higher rates of gun violence (more guns, right?) and “bootstrap pullin’” (more pullin’ up on bootstraps going on there, as well), states with a lot of liberals probably have higher taxes.

Results: Correct! Although, with weak statistical correlation. Call me surprised. I figured the correlation would be stronger, but at a little over 12%, I’m comfortable calling the correlation at least statistically significant.

As it stands, it seems like conservative and liberal states like to tax their citizens in roughly equal amounts. I would be curious as to how these taxes break out (more income taxes vs. property taxes, for instance). On a whole tax basis, though, there’s not much correlation—but the relationship is not negligible, either.

Cost of Living

Prediction: I’m going to go out on a limb and say that more liberal states, with their slightly higher taxes, probably have a small but still significant correlative effect on higher costs of living (defined by BLS as the cost of a basket of goods that correspond with household spending). This would support the justification for that initial Forbes article as to why folks are moving (high housing). What’s more, though, is that urban centers tend to be more liberal and, by nature of being urban, more expensive.

Results: Correct! With strong statistical correlation. The baseline cost of living is 100, so a dot positioned vertically at 120, say, means that state is 20% more expensive than average.

So, it’s official, it’s generally cheaper to live in conservative states that liberal states. I’m not making any arguments here—I’m fairly certain you’d see an equally strong correlation for incomes, as well (meaning, liberal states probably have higher incomes). I just thought this would be a neat chart to see and test my theory.

Corruption

Prediction: My guess is that when more dollars are sloshing around, then you almost certainly must have less oversight over those dollars and at least a greater temptation to be corrupt. Thus, my inclination was that more liberal states probably have greater corruption. But, then I thought for a second, corrupt people are corrupt because they have compromising values—they’re criminals. I don’t think being liberal or conservative inclines you to be a criminal. I’m sure we have some good ol’ boy networks in the deep republican south just like we have had a string of jailed Illinois Governors (four of the last seven, baby!) As such, I’m predicting no correlation between corruption (ranked 1-50 by State Integrity Investigation, with 1 meaning the least corrupt) and political affiliation.

Results: Wrong! A slight statistical correlation. Well, even my initial guess of more liberal= more corrupt, which I backed down from, would have been wrong. As it stands below, you’ll see a slightly upward trendline indicating that more conservative states tend to be slightly more corrupt with a statistical strength of 9.5%.

It’s not a strong correlation, mind you. This lends to my belief that people are either corrupt or not corrupt, regardless of their party. However, you can’t deny that about 10% of a state’s corruption ranking can be attributed to whether it’s strongly conservative. Interesting…

Happiness

Prediction: Okay, this is the ultimate one. What do high taxes and back-dealing politicians matter when your citizens are happy? I mean, being happy is the ultimate goal, right? I’m guessing that people are happy (ranked 1-50 by Time Magazine where 1 = Most Happy) for reasons totally separate from their political affiliation (health, friendships, family, some financial security). So, I am predicting there’s no correlation.

Results: Probably correct, but a weak statistical correlation exists. Statistics can’t explain everything, and this is probably one of them. You’ll see from the chart that there is a 7% or so r-squared upward trendline that favors liberal states. As states become more liberal (going left), they tend to get happier (moving down), but the trend is pretty weak.

Therefore, I’m comfortable calling myself correct on this one. What’s more, if you look at the Time Magazine study that is done yearly to judge a state’s happiness, you’ll see some pretty big movements up and down the list on a year-to-year basis. So, any trend you do see is probably only applicable for a short period of time and, thus, not really significant.

Conclusions

The usual suspects fell on the side one would expect with high costs-of-living, greater tax burdens and worse financial conditions all correlated toward liberalism. Then we saw things like corruption and depression fall against conservatives (who knew!). Then things like interstate migration and happiness are pretty much insignificant.

I’d love to know what you think in the comments below!

Thank You!

-Eric