Thursday, April 4, 2013

Liberal, Conservative, Rural, Urban: A Walk Around the Ideological Space of the 1955-1956 House.


Introduction:

Once again, I ran the "ideal" algorithm, this time on the roll call votes of the 84th House (1955-1956).  The point of these algorithms is to try to summarize all of that information and treat each legislator as a "point", with close points representing clusters of legislators who vote similarly.  Let's see how different issues broke down.

Here's the result:



This time around, I called the vertical axis "urban/rural conservativism", although the overall picture is very similar to my look at the 83rd House.  You can see the "chunk" of (mostly) Southern Democrats seemingly drifting away from the rest of the Democratic party.  But perhaps that's just a modern interpretation.

The algorithm treats each issue or vote as a single line dividing "Yes" votes from "No" votes.  Because Democrats and Republicans are each on their own side, we can make some charts to show different kinds of splits in the House and in each party's caucus.

Obviously, I can't go over every interesting vote, but I will try to give you a sense of which issues created different kinds of "cutting lines".

The Urban/Rural Splits:

As I said above, I think the biggest split, aside from the overall liberal/conservative split, can be interpreted as an urban/rural split.

Here's a good "multidimensional" vote: Roll Call 72, to eliminate a public housing provision from a housing plan.  As you can see, it wasn't too far from a party-line vote, but some (mostly rural and/or conservative Democrats) opposed it, and some (mostly urban and/or liberal) Republicans supported it.





But consider Roll Call 92, a vote on passing an agricultural program over Eisenhower's veto.




This was mostly a party line vote, but a few (urban/moderate) Democrats voted with the Republicans, and a few (rural/conservative) Republicans voted with the Democrats.  Roll Call 28, on farm price supports, had a similar coalition, as you'd expect.

To demonstrate how these two votes reflect the overall dynamics revealed by the algorithm, here's a chart colored with both of them:





The two votes, Roll Calls 72 and 92, split the House into four "clusters":


  1. The blue cluster. Yes on eliminating the public housing provision, yes on the Democratic farm plan. Mostly Southern Democrats, a few rural Republicans like Wint Smith (KS-06) and Usher Burdick (ND-AL).
  2. The dark red cluster. Yes on eliminating the public housing provision, no on the Democratic farm plan.  Mainstream Republicans, including most of the "most conservative" members.
  3. The green cluster.  No on eliminating the public housing provision, no on the Democratic farm plan.  A mix of Republicans and Democrats, mostly, it seems to me, urban and/or northeastern.  Democrats here include Adam Clayton Powell (NY-16), Edna Kelly (NY-10), both Representatives from Rhode Island, and a number from Massachusetts.  Republicans here include Philadelphia's future Senator Hugh Scott (PA-06) as well as representatives from Vermont, New York, New Jersey, and KY-03 (Louisville, then as now).
  4. The black cluster.  No on eliminating the public housing provision, yes on the Democratic farm plan.  Mainstream Democrats.

Of course, any two votes divide a group into (up to) four clusters.  What's interesting here is that all four are relatively large, and are represented in both parties (the green and the blue more than the red or the black).

But here's a third vote on yet another crossing axis: Roll Call 122, to deny school construction aid to states that "refuse to comply" with Supreme Court decisions.  I've seen this referred to as the "Powell Amendment" in one of the CQ Almanacs.





Southern Democrats, unsurprisingly, were unanimously opposed, but so were some fairly liberal Democrats in more rural seats like Coya Knutson (MN-09), of Northern Minnesota; Lester Johnson (WI-09), of Northern Wisconsin (who won in a huge upset in a special election); and future Senator Lee Metcalf (MT-01), of Western Montana.  Of course, this is a broad brush, since some urban liberals like Powell's fellow African-American William Dawson (IL-01), of Chicago, and Martha Griffiths (MI-17), of outer Detroit, opposed as well.

On the other hand, fairly few Representatives from the "urban moderate" cluster opposed the Powell amendment.

Votes Splitting Off Conservatives:

To see the most conservative members, here's Roll Call 61, a conference report on public works appropriations, and Roll Call 63, on extending Social Security benefits.  I assume that the many Pennsylvania Democrats who opposed the former were motivated by some parochial issue.





In the above charts, we can see conservatives like "bitter lone wolf" Clare Hoffman (MI-04), Noah Mason (IL-15), "conservative spokesman", Orange County's James Utt (CA-28--whose district was undergoing enormous growth over the 1950s), and Truman nemesis John Taber (NY-36), who had chaired Appropriations during the Republican Congress.  Aside from Utt, they all also opposed the below civil rights bill.

Votes Splitting Off Liberals:

As for the most liberal Representatives?  Sadly, few individual votes seem to distinguish them.  In the vast majority of the votes with only a few Representatives against everyone else, the few Representatives were from one or another conservative side.

Of course, the algorithms distinguish liberals somehow.  It's just based on multiple overlapping votes, rather than being clear from one or two specific votes.

Some exceptions: Roll Call 80, on disapproving of the government's sale of rubber-producing facilities, and Roll Call 88, on recommiting the appointment of Edwin Howard.  These "sale of rubber facilities" votes are among the clearest individual votes separating "liberal" Democrats from "conservative" Democrats.



As for Howard, apparently, Howard was a Brigadier General who was appointed to an INS position "despite the opposition of some Congressmen who felt the INS was being taken over by the military."



In the above votes, we at least have Richard Bolling (MO-04), later in leadership; Emmanuel Celler (NY-11), Adam Clayton Powell (NY-16), and Sidney Yates (IL-09).

And here's an even more extreme example: Roll Call 143, on citing Arthur Miller for contempt of Congress:


Only nine Representatives were opposed, mostly liberal Democrats (and one Republican, IN-03's Shepard J. Crumpacker Jr., who seems to be covered by adjacent Republicans).  You can see how one of those nine, Abraham Multer of NY-13, might end up as the algorithm's choice for "most liberal" member.

Votes Splitting Off Moderates:

A couple of votes were basically party line, but with some moderate crossover, more or less outside of the urban/rural splits.

For example, here's Roll Call 148, on blocking rate increases by the Southwestern Power Association:



Mostly, Democrats voted against this, and Republicans voted for it, but there are exceptions up and down the party line.

The same applies to Roll Call 112, which was to recommit/kill a water pollution bill:



Other Directions:

For a "horizontal" vote, it's hard to go wrong with a civil rights bill.  Here's Roll Call 136:


This is actually a good way to distinguish the Southerners in that top group from the others.  Districts like OH-06 and MO-11 apparently voted with the Southern districts very often, but not on civil rights.

Compare the related, but distinct, breakdown on Roll Call 60, a vote to approve "mutual security appropriations".  These opposed by most Southerners, but with some moderate Southerners in support, and some conservative Republicans in opposition.  The following chart is colored by both votes.



Of course, some issues are complicated.  Consider Roll Call 49, on funding for the Central Valley Project.  I think the "no" votes are a mix of people saying "why spend all this money in rural California?" (the urban nos) and "why spend any money on anything?" (the conservative nos).  The result is this odd triangle shape.


Which seems as good a place as any to stop.