The Thin Blue Line That Stays Strangely Horizontal

Last night I was pondering how to test the Reactionary claim that society was drifting inexorably to the Left, and I wondered whether maybe someone had created a mathematically rigorous model of people’s positions along the left-right spectrum and then applied it to every legislator in every session of Congress from the founding of the United States to the present day.

This being the Information Age, of course the answer is yes. NOMINATE is a statistical system that takes every legislator’s vote on every issue that comes before Congress, then number-crunches it to see which legislators are closest or furthest away from one another. The classic left-right dimension naturally falls out of such a model, allowing a quasi-objective way to determine which legislators are more liberal or conservative than others.

It would appear that this could only compare two legislators serving at the same time and voting on the same issues. But an enhanced version of the system, DW-NOMINATE, gets around this limitation with a clever trick. Political scientists know that legislators’ views generally change little over time. Therefore, long-term legislators can be used to “bridge” different sections of Congress. If Senator Bob is the most liberal Senator in 1950, but by 1970 he’s somewhere in the middle, then we know the Senate became more liberal from 1950 to 1970. If Senator Mary is the most liberal Senator in 1970, but by 1990 she’s somewhere in the middle, then we know the Senate also became more liberal from 1970 to 1990. Therefore, we can conclude that the Senate in 1990 is more liberal than in 1950, even if no single Senator served in both bodies.

DW-NOMINATE has been called “one of the great achievements of modern political science”, has won a bunch of awards, and most tellingly is routinely used by Nate Silver. It is recognized as broadly accurate by people from both sides of the political spectrum. So let’s see what it finds:

A few things here stand out as correct. Congress is getting more polarized recently – true. The Republicans took a big turn to the right starting in 1980 – true.

But from this it looks like there has been zero liberal drift of either party over the past 120 years. And don’t get me wrong, if it’s true I will relish rubbing it in the Reactionaries’ faces – but it seems kind of implausible. That whole “expansion of the welfare state” was a thing that happened, right? And Martin Luther King Jr speeches aside, I’m pretty sure the civil rights movement wasn’t literally a dream. Gay marriage? Also, either we got an income tax early in the 20th century, or else some guy with the initials I.R.S. has been pulling a very successful con on me.

Normally I’d dismiss this. I’d say it probably wasn’t kosher to use DW-NOMINATE data to make time comparisons. But a political analyst tried making that argument last year and was shot down by the developers of the DW-NOMINATE system. And Nate Silver seems pretty happy to draw DW-NOMINATE comparisons across presidents. And from what I understand of the long but gratifyingly non-mathematical guide to using DW-NOMINATE, intertemporal comparisons are entirely acceptable as long as they don’t go back further than 1879. And the system’s ability to pick up other secular trends, even smallish ones, seems quite good.

I’ve got to admit it. I’m stumped on this one.

Edit: There’s a similar measure called Martin-Quinn scores for the Supreme Court that finds a similar result

This entry was posted in Uncategorized and tagged , . Bookmark the permalink.

50 Responses to The Thin Blue Line That Stays Strangely Horizontal

  1. Douglas Knight says:

    1. Political scientists know that legislators’ views change very little over time.

    This is a key assumption for trusting this methodology. Moreover, I think that whatever method establishes this claim ought to be able to address the main question. So I think you should pursue this and not use dw-nominate to answer this question.

    2. A decade ago I saw a similar technique, probably by the same people, that assigned left-right as the first principal component and civil rights / slavery as the second. So by that measure, a shift in civil rights isn’t a shift to the left.

    • Scott Alexander says:

      I’ve added to the post a link to Poole explaining why he thinks legislators’ views change little over their careers: http://www.voteview.org/chminds.pdf

      • Douglas Knight says:

        It seems pretty important to me it’s the same guy with the same methods drawing this conclusion.

        It appears circular to me. According to this blog post you linked, dw-nominate (1997) assumes that individuals are stable in order to compare different years. They justify this assumption by invoking the paper (2007) you link, which uses dw-nominate. (It also uses other methods, but they’re all about the same.)

        All that I can see you can conclude from this kind of data is that politicians who are extreme while young are extreme while old; and those that are moderate while young remain moderate. If that is what they’re concluding, it is interesting and it rules out certain popular models, that individuals move right with age or that that individuals stay put and the population moves left.

    • Vaniver says:

      This is a key assumption for trusting this methodology. Moreover, I think that whatever method establishes this claim ought to be able to address the main question. So I think you should pursue this and not use dw-nominate to answer this question.

      They give people different coordinates in different Congresses, and fit a linear trend through time. I suspect it works okay.

      A decade ago I saw a similar technique, probably by the same people, that assigned left-right as the first principal component and civil rights / slavery as the second. So by that measure, a shift in civil rights isn’t a shift to the left.

      I looked into this directly, and it appears this is correct (i.e. when the parties shift on civil rights, the Left/Rightness of civil rights issues changes), which seems to seriously reduce its usefulness as an anti-reactionary argument and intertemporal comparison.

      • Douglas Knight says:

        They allow people to drift at constant velocity with time, but everything is relative. If everyone is drifting at the same speed, their model can’t detect it. Indeed, abstract role call votes can’t detect it.

        • Scott Alexander says:

          For this to work, individual legislators would have to be drifting at just the same rate as society as a whole – ie there would have to be no “politics advances funeral by funeral” effect where young people are more plastic to new political ideas than older people. This seems to violate both evidence and common sense.

        • Fnord says:

          Is individual legislators drifting at just the same rate as society as a whole less plausible than no drift at all?

        • Douglas Knight says:

          For what to work?

          Poole says that this data says that individual incumbents are drifting at the same rate as the drift due to turnover, where both drifts are averaged over the whole period. It says that regardless of what you assume the drift is. If you reject that, you should be concerned about your ability to extract any long-term conclusion from this kind of data.

  2. suntzuanime says:

    Legislators’ views change little over time? http://en.wikipedia.org/wiki/Robert_C._Byrd

    Maybe you could argue that his views actually did change little over time, but certainly those of his views he felt were wise to express headed steadily to the left. I don’t think DW-NOMINATE, for all its coolness, is capable of picking up secular trends such as a shifting Overton Window. Anything that affects all legislators at once is going to be unnoticeable.

  3. nazgulnarsil says:

    If the leftward drift is basically baked in, and thus overlooks all the things you mentioned in your second to last paragraph, then all DW-NOMINATE is saying is that how far the left and right diverge from the slowly-drifting-leftward center hasn’t changed too much.

  4. Brian says:

    So xkcd did a comic about this. I think the big factor is that “what is conservative” (or liberal) over time, shifts. See the idea of the “Overton Window” and how people can force it to shift.

  5. Vaniver says:

    I think you’ve got a typo, since you saw DW-NOMINATE is an improvement on DW-NOMINATE. The differences between NOMINATE, D-NOMINATE, W-NOMINATE, and DW-NOMINATE is explained here.

    I wonder to what extent this is due to looking at party means. If we looked at each member of Congress as a dot each term, maybe colored by their party membership, would we see more? For example, when we look at just party means, this hides shifts in the percentage elected of each party. Looking at the Congress mean would solve that.

    I find the Southern Democrat flip suspicious, because the Southern Democrats of 1895 were the party of “Pitchfork Ben” Tillman, open white supremacist. From Wikipedia:

    In 1901, after President Theodore Roosevelt dined in the White house with Booker T. Washington, Senator Tillman said, “The action of President Roosevelt in entertaining that nigger will necessitate our killing a thousand niggers in the South before they learn their place again.”

    DW-NOMINATE puts him at (-0.367, -0.064), and he’s only 28th out of the 100 senators (with ‘most liberal’ at 1) in the 54th Congress.

    The 28th most liberal senator in the 112th Congress is the late Daniel Inouye, big damn hero who the NAACP gives an 89% rating. Who’s the “closest” to Tillman? They say to not use the Euclidean weighting, but don’t say what weighting I should use instead, so I’ll use it anyway. Menendez with a (-0.392, -0.101) (17th most liberal) narrowly beats Mikulski, who has a (-0.395, -0.100) (14th most liberal). Both of them got a 100% score on the NAACP’s legislative report card.

    African American rights was literally the founding issue of the Republican party, and unsurprisingly the Republicans had near-total support from blacks. Now, the party that has near-total support from blacks is the Democratic party, and the Republicans are constantly accused of being opposed to blacks.

    (I was about to type “But to the extent that DW-NOMINATE obscures the fact that senators no longer openly call for the murder of American citizens,” but then I realized that I would need to put a lot of qualifiers on that sentence. So scratch that.) To the extent that race is a huge political issue in the US, it seems that DW-NOMINATE is inadequate to capture long-term shifts of the sort that reactionaries think are important.

    • Scott Alexander says:

      I’m uncertain how to deal with civil rights in this system. They mention that their model does sometimes have a second dimension that sort of corresponds to civil rights, but it looks like also the second dimension can correspond to random other things at any time like bimetallism, and also that the second dimension sometimes disappears entirely. I also think I saw it mentioned that the second dimension only explains about 2% of the variation beyond what the first dimension explains.

      I wonder what would happen if they just tried to force everyone into one dimension.

      • Vaniver says:

        Civil rights isn’t one of the two dimensions- if it were, then Tillman and Mikulksi would be very different on that dimension. The second dimension is probably something that the two parties are generally equally split on.

        I think that bimetallism is a good example of how these comparisons probably don’t argue against the neoreactionary view. For anyone not familiar with it, here’s a basic intro.

        If you view it as “owners” versus “borrowers,” then you can make comparisons to today by asking which contemporary issues favor owners more, and which favor borrowers more, and how the various parties behave on those issues. You probably see lots of continuity this way, because owners and borrowers are always opposed and there’s always a party that goes for owners’ votes and a party that goes for borrowers’ votes.

        If you view them as literal positions, though, then basically the only Congressman on the 1895 political map from the 112th Congress is Ron Paul, widely seen as a kook today, and both the Republican and Democratic party means are totally off the charts. This is what the neoreactionary thinks is interesting, and I think most issues will play out like this. Do any issues come to mind where the Overton window hasn’t significantly changed over the last 150 years?

    • Damien says:

      Yeah, while I can feel the chart accurately measures the distance between e.g. Republicans and Southern Democrats across time, I don’t think it’s accurately portraying those in relation to the modern day positions. There should be some sort of twist in at least one more dimension.

      I’ve read the 1856 platforms. The Republican party was the anti-slavery party, but also one of big business, always. The Democrats were pro-slavery, especially in the Southern wing, and against “internal improvements”, but they were also pro-immigrant, and pro-Catholic rights. They’ve always been the populist party, especially in the north.

      But even so, 1880s Southern Democrats to the “left” of northern ones? I roll to disbelieve.

  6. John H says:

    How is the liberal/conservative spectrum measured?
    If it’s how the politicians define themselves, you’re unlikely to see a long term shift as conservative/liberal is usually defined as one’s distance from the center; the actual positions of liberals and conservatives have changed dramatically over time.

    So I guess what this metric would do well is see how far the average politician distances himself from the mean over time (which is why the increasing polarization shows up), but not whether society has shifted to the left or not.

    Political ideologies being as complex and arbitrary as they are, it’d be hard to quantify a shift ‘to the left’. Tax rates advocated by either party are enormously lower than they were pre-Reagan, but at the same time you have increasing support for womens’ rights and interracial marriages.

  7. Damien says:

    “But from this it looks like there has been zero liberal drift of either party over the past 120 years.”

    Well, the curves don’t look exciting, but paying attention to the numbers, the GOP went from 0.5 in 1903 to 0.2 in 1975, before soaring to 0.7 today. There was a long slow shift, more recently and suddenly reversed (and then some)

    Northern Democrats are more “liberal” than they ever have been, at -0.4, vs. starting point of about -0.3 and peak of -0.2. The weirdness of the Southern Democrat curve has already been mentioned, but heads left regularly from 1945… judging by how Democrats converges with Northern Democrats, I think the graph implies Southern Democrats have mostly passed out of existence.

    True, if you just compared the endpoints, the Democrats are right where they started and the GOP has headed right, but the history is a lot more complex.

    …but yeah, when we’ve seen changes on issues like black vote, women’s vote, income tax, welfare, Social Security, gay rights, etc., the graph looks odd.

    OTOH, if you just looked at economic issues, unions and inflation and bank regulation and top income tax and universal health care (votes, not passed bills) then the graph might be a lot more plausible, or even portray the Democrats as more liberal than they really are. But if it’s only or mostly economic issues then it’s probably not measuring what the reactionaries complain about.

  8. Douglas Knight says:

    I’m 70% confident I’m not making a basic error about the intent of the statistic, where the people who invented it would object “No, you shouldn’t even try to use our method to measure that.”

    Yes, they’d probably woudn’t say that. You found several places where they say the opposite. But I say it.

    It is simply not possible to extract an overall trend from a time series of abstract roll call votes. If the questions being voted on drift at the same speed as the voters, the distribution of votes is the same every year. There is no way to figure out how fast the drift.

  9. lmm says:

    Does the axis measure what you think it does? I mean, are scores on this measure correlated with voting records on expansion of the welfare state, civil rights, and Gay marriage? Or is it e.g. overwhelmingly a measure of support for taxation, with all other views getting swamped by that?

    What do representatives who switched parties look like on this measure? (Is that a thing that happens in the US?) Are they people whose views jumped, or who found that a different party was now a better match for their static views? Is it more common to switch from a leftist party to a rightist or vice versa?

    • AUroch says:

      >is that a thing that happens in the US?

      Only once or twice a decade, at the national level, and they usually lose their next election.

  10. Erik says:

    I feel like I’ve walked into a debate where some people are considering the possibility of Antichthon or Counter-Earth, a planet matching Earth’s orbit but hidden on the other side of the Sun (Gor being a fictional example of such).
    An astronomer happens to overhear part of their conversation and contributes: “I was on the other side of the Sun six months ago, and there was definitely a planet there at the time.”
    Then the astronomer adds: “Fairly sure there’s nothing there now, though.”

    That sort of describes how confused I am here and how much I think this is talking past one another/measuring something entirely different.

  11. DW-Nominate scores politicians relative to each other. Therefore if *everyone* becomes more liberal at once (both parties), there won’t be any visible drift in the graph.

  12. Crimson Wool says:

    Here‘s a version which is the House/Senate on one graph, no party identification. It makes it clear that there is, in fact, variation over time, so “everyone co-varies” can’t explain things alone. Perhaps liberals are more effective at long-term policy shifts? After all, I can think of very few “reversions” of major laws (Prohibition is the only one that comes to mind), but many liberal laws/ideas that remain in place. So maybe conservatives fight really long and hard against, say, desegregation, and then after desegregation is done they virtually spend no time trying to get it put back in place. If it reappeared by some mystic action by the Reactionary God, they’d fight to preserve it again, but they don’t fight to re-enact it. The same would apply for the welfare state, gay marriage (presumably), etc. They fight and fight and fight and eventually the law gets passed and they don’t fight to get it repealed (they do fight, actually, but not for very long).

    • Scott Alexander says:

      Thank you. That graph seems much more to correspond to political changes than mine (for example, it visualizes the liberalism of the LBJ years). Where did you get it?

      Your idea about long-term policy shifts makes a lot of sense and in fact was something that got brought up in the Twitter commentary to this post last night.

  13. I’m as confused about this as you are, Scott, but I have a few thoughts anyway:

    1) This does seem to fit with what some people’s un-rigorous observations on the political direction of the US. E.g. supposedly Reagan, in spite of being revered by present-day Republicans as the great tax cutter, was more flexible on taxes than present-day Republicans.

    2) So it’s possible leftward drift on some easy-to-understand issues has been more than balanced by rightward drift on other issues. Top income tax rates went from 70% when Reagan took office to 39.5% today, and would probably be less if the Bush tax cuts hadn’t had a sunshine clause. Though harder to measure, regulation seems like another good candidate for an issue we may have drifted right on.

    2a) You’d have thought 9/11 meant a rightward jolt on war and civil liberties, but that doesn’t seem to be reflected in the graph.

    3) I wonder if the graph partly reflects things that don’t actually affect what legislation gets passed, such as:
    * Post-2010 House Republicans constantly voting to do things Obama and the Senate will never go along with, Democrats not having done anything similar at any point
    * (Willingness of Democrats to go along on bills that are primarily driven by Republicans) > (Willingness of Republicans to do the reverse)

  14. Alexander D says:

    I would suggest that the spates of progress, when compared with the stable liberality, are artifacts of the system itself. For many, many years, the Southern conservative bloc stopped an enormous amount of change from getting through Congress. The New Deal and atmosphere of the time swept through the House and scared the Senate (although it didn’t change it that much) and so a lot of legislation that had been blocked suddenly went through, as the minority stopped its stranglehold. Then again, the same thing happened once Lyndon Johnson, as Senate Majority Leader, eroded the power of the Southern conservative committee chairmen, using the institutional changes and his own affiliations with those leaders to push through some legislation (as well as his own ambitions). With those chairmen’s power reduced and the leader’s power increased, other liberal legislation was passed and even conservative Presidents had to deal with a generally liberal Congress for many years (Clean Water, Clean Air, etc.)

    You’re looking at institutional changes, because DW-NOMINATE reflects relative ideology, and that ideology is filtered through the institutions before it becomes law.

  15. Mary says:

    And Martin Luther King Jr speeches aside, I’m pretty sure the civil rights movement wasn’t literally a dream.

    Yes, but it was reactionary, not progressive. It was the progressives who swept to power and gave us segregation; Woodrow Wilson’s election was the cue, as he stopped stomping on southern states who implemented it and instead went and segregated the federal workplaces.

    Since King et. al. wanted to return to the practices of the Reconstruction, they were reactionaries.

    • Charlie says:

      For the end of Reconstruction are you confusing Wilson with Hayes, 40 years earlier? And if both sides in a dichotomy have historical precedent, what is the worth of saying something is reactionary because it has historical precedent?

        • Charlie says:

          Fair enough, I guess there is a perfectly reasonable interpretation where “It was the progressives who swept to power and gave us segregation” means that at some point, the progressives (Woodrow Wilson Democrats) swept to power and increased the amount of racism in the US government.

          Rather than the interpretation that actually has any bearing on what is later argued, where “It was the progressives who swept to power and gave us segregation” means that segregation was the progressives’ fault.

        • Damien says:

          Well, Wilson didn’t invent segregation, but he did bring it to the federal government and military.

          Ah, but Mary said “It was the progressives who swept to power and gave us segregation”. Yes, that’s ahistorical BS.

          As for “progressives”, I liked the features Scott gave Raikoth’s putative conlang, starting with “no unquantified plurals”. So we need something like ‘all’, ‘most’, ‘some’ etc. in front of progressives. Along with “what is my confidence level” and “what is my evidentiary basis”.

  16. Ishmael says:

    People who live farther from big cities have more right-wing opinions, and are over-represented in the legislature. North Dakota has had a population of about 1/2 million for a hundred years, and 3 representatives for all that time. Over the same 100 years the population of Massachusetts grew from 3 1/2 to 6 1/2 million, while its number of federal legislators decreased from 15 to 12.

    So you should certainly not use NOMINATE as a proxy for something like “how far to the left or right the average voter is.”

    Another mystery, why does the government seem more left-wing now than a hundred years ago, if the legislators opinions are about the same now as ever? If you’re a reactionary, you have an easy explanation for this: legislators don’t make policy.

    • Charlie says:

      Hm, except that one could replace “policy” with “statements about gay marriage” and retain the apparent liberal shift.

      I think it’s more likely that NOMINATE is at fault, possibly by grouping too many things together – every time the republicans and the democrats split a vote over whether to name a new bridge after Ronald Reagan, NOMINATE uses that data to group them. And every time they vote for against the Civil Rights Act, NOMINATE uses that data too. But there are a lot more bridge-namings than Civil Rights Acts.

      Though then again, maybe NOMINATE just screws up in its foundational assumption that people don’t shift their policy votes very much over their lives. The Civil Rights Act couldn’t have passed before the Civil Rights movement really got into full swing. But there wasn’t that much turnover in the Senate between “unpassable” and “passable” – there was just a change in the surrounding political climate.

  17. Fnord says:

    Having finally followed the link to read the Nate Silver article, I wouldn’t exactly call it a ringing endorsement of DW-NOMINATE’s ability to compare across time.

    I’ve given a lot of credence to the DW-Nominate scores in this article, assuming that they do a reasonable job of capturing shifts in the ideological positions of Congress over time. In truth, after having read “Ideology and Congress,” I’m not entirely persuaded that they can capture all of these dynamics. The system is essentially blind to the content of legislation, so if there are changes in the types of bills that Congress votes upon, there could be long-run ideological changes that are not well accounted for by the system. Measuring ideological change is one of the trickiest questions that political scientists face, and a complete treatment would require a thesis- or book-length approach.

    And that’s the caveat he gives after using DW-NOMINATE to compare across a mere 20 years, not the 140 we’re talking about here.

  18. Tyrrell McAllister says:

    How does the following hypothesis stand up? Maybe the big “liberal” developments just haven’t translated into all that many votes.

    Consider desegregation. This was an issue that dominated public discussion. We intuitively recognize it as representing a sea-change in the nature of American society. But did it come up for congressional vote all that often? Or were there just a few super-high-profile bills where congressional votes correlated strongly with opinions on desegregation? Such opinions might have been hugely influential on society at large while influencing only a hand-full of congressional votes.

    I could imagine a similar story about gender roles, abortion, divorce, and sex in the media. Maybe the social issues on which conservatives feel like they’ve lost the most ground just don’t come up for a vote that often. There might be a few high-profile votes, but the matter is mostly settled in the broader society, not in Congress. Similarly with the income tax. It might have been a major issue, but how many votes depended on it?

    I’m just throwing this story out there. I haven’t subjected it to strong critique myself, yet. But if it’s right, wouldn’t the few votes representing the biggest “liberal” disjunctures with the past have been completely overwhelmed in the DW-NOMINATE data by the myriads of votes on minor economic and regulatory issues surrounding things like, say, agricultural tariffs?

    A 1900’s “liberal” like Tillman might have been completely off-the-walls crazy by modern liberal standards, but how many of his votes were so at odds with those of a modern liberal senator?

    • Deiseach says:

      “Maybe the social issues on which conservatives feel like they’ve lost the most ground just don’t come up for a vote that often.”

      That seems plausible; most of the changes in law seem to come about not because of a bill introduced in parliament but through the courts (e.g. Roe vs. Wade).

      What about parties/members of parliament who are socially liberal but fiscally conservative – do these come across as skewing left or right? From Irish politics, we had one party (since disintegrated) which was very hard right on Big Business but much more towards the left on divorce, contraception, etc. Mostly, though, our political parties all jostle to be the most ‘centrist’, so former ‘core principles’ get jettisoned if it looks like that will win more votes.

      Our current government has one of the two major political parties in charge; they were always regarded as right wing, and remain ‘right’ on business etc. but have moved dramatically to the ‘left’ on such issues as – for the first time in the history of the state – legislation permitting limited legal abortion.

      So how would this measure grade Joe Murphy, T.D. for East West Midlands, who votes “yes” on the Zero Corporate Tax Bill and also votes in favour of the Gay Marriage Now bill – right or left?

    • rrb says:

      I think this could be it. Is it checkable? I hope someone checks it somehow.

  19. Eric Rall says:

    One possibility is that what’s largely static in a politician over the course of his career isn’t necessarily specific policy positions, but rather his temperment and factional affiliation, which combine to produce a comfort zone relative to the overton window. For example, Reagan as a political activist in the 1960s strongly opposed Medicare and Medicaid, but as President in the 1980s made no major effort to arrange their repeal — the overton window had shifted, and Reagan’s policy goals had shifted with it. He was still aligned with the conservative wing of the Republican party, and he still favored policies that were about 3/4 of the way towards the right edge of the overton window (number guestimated), but the specific policy mix had changed with the times.

    Under this hypothesis, the shift in the DW-NOMINATE graph represents changing factional composition within each congressional party caucus. When one faction turns out another as the dominant force in a party’s congressional caucus, that would get captured very well by the DW-NOMINATE methodology. But if each faction’s positions shift together because of exterior factors (shifts in the overton window), then that would get missed.

    In addition to the anecdotal evidence offered by other commentors, I’d also like to offer the trajectory of government spending as a % of GDP as evidence of a divergence between policy outcomes and DW-NOMINATE scores: (graph)

    The immediate appears to have an axe to grind, but he cites some good neutral sources for his data, and I don’t see obvious problems with how he composed the graph.

    • Douglas Knight says:

      One possibility is that what’s largely static in a politician over the course of his career isn’t necessarily specific policy positions, but rather his temperment and factional affiliation, which combine to produce a comfort zone relative to the overton window.

      Yes, this is the kind of question that this kind of data could answer. Poole claims that this is what happens. But Scott doesn’t believe even that claim.

  20. Tom Hunt says:

    One might also suggest that the positions of politicians are measured in, essentially, the derivative of the “societal position” value that seems to have drifted significantly leftward over the past century. Thus, shifts in legislators’ positions, or even party averages, don’t significantly change the calculus so long as the overall average (which is a first-differential value) remains positive. You’ll notice that, in mainstream political discourse, Republicans may staunchly resist the passage of liberal bills which are newly introduced, but not nearly as much effort goes into attempting to undo past changes. They may quite strongly resist passing new gun control laws, but no one is seriously proposing an effort to repeal NFA ’34 or GCA ’68. They may oppose new entitlements, but no one serious is trying to actually do away with Social Security. And so on.

    The reason why it’s liberal causes, rather than conservative ones, which display this ratchet effect is a different one, and probably deeply tied in with neo-reactionary philosophy somehow.

  21. Erik says:

    I’m still confused. I don’t understand the paper arguing that legislators’ views change little over time. It seems to assume much of the conclusion. At the top of part 2 (part 1 being the introduction) we get this:

    “I make the assumption that all members of Congress who do not change their party affiliation including those who change from the House to the Senate have fixed ideological positions throughout their careers.”

    What.

    • Douglas Knight says:

      Vaniver said that they allow people to move through ideology space at constant speed. After I read that, I thought I saw it in the paper, but now I can’t find it.

      Maybe what he means by that sentence is that one model makes the assumption and he compares it to another model, but I can’t tell what that other model is.

  22. Douglas Knight says:

    It’s funny that they say you should only use their model with two parties, but then they say that you should think of the Southern Democrats as a third party. If the concern is separating parties, two dimensions are enough to separate three parties.

    But they don’t say two parties, but the two-party system. I’m not sure why the two-party system specifically matters, but the party system has a huge effect. Parties exist to negotiate and compromise on an ideology. Maybe the ideology of a party is the average of those of its members, but the party greatly decreases the variance. People subordinate their views to the party; that is what they contribute to the party. Poole quietly admits this when he says that people who switch parties will jump in ideology space, and thus he counts them as two people. So of course this method will draw emphasis to the axis between the two parties. That one axis explains most of the variance is an artifact of having two strong parties. But it’s still probably a good proxy for the left-right axis. I don’t have any conclusions to draw from this comment, but I think it’s worth keeping in mind.

  23. Pingback: The Communists Won | Free Northerner