Southern Election Data

(I divided the spreadsheet so it would fit comfortably on your screen.)

So I got curious about trends in the Southern election data, (see yesterday’s post on Northern election data and last week’s post about my migration/Civil War theory,) thinking to myself that perhaps an opposite trend happened in the South–maybe poor sods who couldn’t catch a break in slavery-dominated states decided to go test their luck on the frontier, leaving behind a remnant population of pro-slavery voters.

Methodology/discussion:

I took as the “South” all of the states south of the Mason-Dixon. This turned out to be incorrect for Delaware and Maryland, which both tended to vote against the Southern states; Delaware, IIRC, voted with Massachusetts more often than “Northern” New Jersey.

The practice of having the legislators rather than citizens vote for president persisted for longer in the South than in the North, especially in SC, which did not have popular voting until after the Civil War; all of SC’s votes here, therefore, come from the legislature.

A “yes” vote means the state voted with the Southern Block during the age before individual vote counts were recorded or the state did not allow individual voting. A “no” vote means the state voted against the Southern Block under the same circumstances.

Originally I had planned on using VA as my touchstone for determining the “Southern” candidates, but VA did not always vote with the rest of the South. So I decided which candidates were the “Southern” ones based primarily on how badly they polled in MA.

A few of the elections had some weird anomalies.

Four candidates ran in the 1824 election. Only one of them was popular in NE, so that was easy, but the other three each won electors in the South, which resulted in the election being decided by the House of Representatives. In this case, Jackson carried most of the Southern states, but not VA or KY, so I decided to count only votes for Jackson.

In 1832, SC decided to cast all of its votes for the “Nullification” (State’s Rights) party. Since “States Rights” is the more polite form of Civil War grievances, I decided to count this as SC voting in line with pro-slavery interests, even though it was not in line with the other Southern states.

In 28 and 32, the states of Georgia, Tennessee, Mississippi, and Alabama seem unsure how this “voting” thing works, and returned unanimous or near votes for their chosen candidates. Many Northern states also had anomalously high percents in those yeas, IIRC, so this may not be voter fraud so much as everyone just feeling like they ought to vote for the same guy.

In 1836, the Whigs ran four candidates in hopes of throwing the election to the House again, resulting in a fragmented Southern block. I counted all Whig candidates as part of the MA/Puritan side, and so give here the vote percents for Van Buren, the Democratic candidate.

In 1856, the Whig party had disintegrated, and two parties took its place. The Republicans, soon to be very famously anti-slavery, emerged in the North but do not appear to have run at all in the South; I don’t think they were even on the Southern ballots. In the South, an anti-immigrant/nativist party sprang up to balance the Democrats. It won few states, but performed well overall. I couldn’t decide whether to count the Democrats or the nativists as the more pro-South / pro-slavery party, so I wrote down both %s, Dems first and then nativists.

This oddity persists in 1860, when again the Republicans do not appear to have even been on the Southern ballots. The Democrats split in two, with one candidate running in the North against Lincoln, and another candidate running in the South on an explicitly pro-slavery platform, against the the “pro-union” party whose main platform was opposing the civil war. The Union party polled decently throughout the South–taking VA, KY, and Tenn.–but received very low %s in the North. The North, it appears, was not as concerned with trying to stop the Civil War as Virginia was.

Conclusions:

The data does not support my suspicion that less-slavery-minded people moved out of the Southern states. In fact, the most ardently pro-slavery, pro-secession states were Mississippi, Alabama, Arkansas, Florida, and Texas, who also happen to be the last 5 Southern states admitted to the Union, with last but not least Texas outstripping them all at 75%. In that same election, Virginia, the first Southern state, voted for the pro-union party.

So it looks like the same pattern appears here as in the Northern data: more conservative people have moved Westward.

However, the %s voting for the Southern candidates held fairly steady once the era of unanimous voting ended. Georgia, for example, went from 48% 1836 in to 49% in 1860. Mississippi went from 59% to 59%. VA hovered around 55%-50% until the last election. So I don’t see any clear trend of coastal states becoming more liberal over time, aside from maybe VA.

Pre-Civil War Election Results by State

Simplified version

Detailed version

I wanted to see if I could find any evidence in support of the theory I discussed the other day that emigration of more “conservative” folks from the Northern East Coast toward the American West left behind an increasingly liberal population that became increasingly concerned about slavery.

I decided to use election results as the easiest way to gauge relative “liberal”ness over time. If the theory is correct, we ought to see an increase in the % of the vote going to “liberal” candidates over time on the Northern East Coast, and a lower % of the vote going to “liberal” candidates in the western areas settled by former East-Coasters.

Since the theory does not concern itself with the behavior of Southern voters, I ignored them completely and only looked at election results for states north of the Mason-Dixon.

Ideally, I’d also look at Congressional elections, but I decided that focusing on Presidential elections would let me quickly get a general idea of whether or not I had a potentially viable idea.

Some notes:

The first difficulty in compiling this data was deciding who the “liberal” candidates were. I eventually decided to dispense with the notion entirely (see Friday’s grumbling on the subject of whether the Puritans or Jamestown settlers were more “liberal.”) Rather, there is a very clear pattern in the data of Massachusetts and Virginia voting for different parties; MA and the other Puritan states tend to vote together, while VA and the rest of the South tend to vote together. Taken over the long haul, the voting pattern looks more like two ethnically-based parties than two ideologically based parties. Since I am not interested here in some Platonic ideal of “liberalness”, but merely the ideas that prompted the Northern attitudes toward the Civil War, I decided to regard whichever party was dominant in MA (the most consistently anti-whoever-won-VA state) as the “liberals.”

I’ve ordered the states by year of settlement for the original colonies and year of entry to the US for the rest.

There’s no data on popular votes for most states prior to 1824, mostly because most states didn’t have popular voting back then. These elections I have marked merely “Yes” or “No” for “voted for the Puritan candidate” or “voted against the Puritan candidate.”

There’s a period of mass agreement that affected most–but not quite all–of the states until about 1832. The election of 1820, for example, was nearly unanimous–only one elector chose to vote against James Monroe. This I suspect has more to do with the whole country being relatively novel (and early elections lacking popular voting,) rather than mass Puritan or anti-Puritan sentiment.

Starting in 1840, third parties with anti-slavery platforms appeared on the scene and quickly grew in significance. These parties polled pretty much zero outside of the North. The “detailed” dataset gives both the main party’s total and the third party’s totals; the “simplified” set only shows the composite total. (Since it I’ve been working on this very late at night, I hope the math is all correct, but if you find an error, I’d appreciate knowing about it.) I regard the emergence of these third parties as evidence of further leftward movement of the voting public.

The third party polled particularly well in 1848 due to having a very popular candidate; this should not be seen as evidence that the party itself was as popular as it looks in ’48. Likewise, 1836 had a very unpopular candidate.

Interestingly, this period also saw the emergence of a very small fourth party in the North devoted to opposing immigration. I regard these as local “conservative” parties and so didn’t include them in the graph.

Several states are “border states” that received significant migration from both the North and the South; they should be considered accordingly.

Conclusions:

The data looks tentatively favorable to the theory.

If we ignore the period of mass agreement, MA’s support of the Puritan candidate goes from 47% to 62% between 1832 and 1860.

NY holds fairly steady (perhaps because NY is a large enough state that many of its migrants stayed within the state,) but increases from 48% to 54%.

RI: 50% to 61%

Conn holds mostly steady, but increases a bit, from 55% to 58%.

NH: 43% to 57%

New Jersey, a border state, went from 49% to 48%. (Though it was 42% in 1824.)

Pennsylvania, interestingly, did not vote with MA at the beginning, but consistently voted against it. In 1832, it appears that the Puritan candidate wasn’t even on the Pennsylvania ballot. Generously ignoring 1832, Penn makes a remarkable rise from 12% in 1824 to 56% in 1860.

VT: 35% to 76%

Ohio did not originally vote with the North; it went from 25% in 1824 to 52% in 1860.

Indi: 19% in 1824 to 51% in 1860.

Il: 32% in 1824 to 51% in 1860.

ME, formerly part of MA, follows the general coastal New England pattern of mass agreement in the early yeas, then goes from 44% in 1832 to 62% in 1860.

Missouri also seems to have not run the Puritan candidate in 1832, but otherwise went from 4% in 1824 to 44% in 1852, then dropped down to 0% and 10% in the final two elections–most likely due to confusion and campaign difficulties after the Whig party dissolved and the Republicans took their place, rather than a sudden massive shift in attitudes.

Michigan went from 46% in 1836 to 57% in 1860.

Iowa: 50% in 1848 to 54% in 1860.

Wisconsin: 62% in 1848, then down to 57% in 1860.

CA: 47% in 1852, then down to 42 % in 1860.

Minn: 63%

OR: 36%

In the 1860 election, the Puritan candidate polled above 60% in MA, RI, VT, ME, and Minn. Four of those are old Puritan states, and I think Minn is full of Scandinavians.

Puritans polled below 55% in NY (those Dutch!) Ohio, Indiana, Illinois, and Iowa (all Midwest states,) below 50% in NJ (border state,) and Missouri (both a border state and probably a fluke,) and below 40% in the most Western states, CA and OR.

Thus I conclude that we see a general trend in most of the Eastern states toward increasing support for the Puritan candidate, while the more Western states, despite their Puritan transplants, showed much less (sometimes decreasing) support.