January 30, 2017 § Leave a comment
I hadn’t been intending to write this post, but my twitter feed has been flooded with comments all day on this issue after I commented on a tweet of Nate Silver’s about it; voting percentages, that is to say, the % of the time a legislator votes with a certain leader or leader’s position. How useful are they, what can they tell us, and how do they tell us that?
Nate has set up a section of his (fantastic and worth reading) FiveThirtyEight website to deal with the question of how much certain legislators do or do not vote alongside Trump. You can find the figures here. He also gave a (presumably sarcastic) congratulation to CA Senator Feinstein for her place at the top of the chart, with a surprising 100% vote-with-Trump score despite a very liberal constituency. Looking at the data more closely, however, something is very obvious – there haven’t been many votes so far since Trump’s inauguration. So how much should Feinstein be singled out? Does that 100% score represent a unique betrayal of her fairly leftwing constituents? How different actually is her voting record to other senior Democrats?
I will add a further disclaimer before answering these questions – my intention, when I queried the wisdom of Nate’s singling out Feinstein, was decidedly not to defend any of her specific votes. It is already retrospectively clear that any Democrat backing John Kelly made a poor choice and one they should be held to account for; Mattis may be the most defensible of Feinstein’s votes, though he seems likely to be sidelined in the administration (and it’s not a defense I’d like to make myself!). Many enraged leftwing Democrats and other US leftists assumed nonetheless that defending Senator Feinstein was my intention and piled onto me, including in one case my being accused of “using logic to enable fascism”, which is a new one. My only point was to discuss whether a table of percentile figures was, in this case, reasonable and whether singling out Feinstein as uniquely pro-Trump for a Democrat made sense.
The answer is; no, it wasn’t, and no, it doesn’t. The Senate has made six votes so far, Feinstein has voted in five of them. The one she missed thanks to illness is the only one that’s gone on dead-partisan lines so far, regarding the budget for the Affordable Care Act. She voted in favour of James Mattis as defence secretary (2 votes), and Mike Pompeo as CIA Director, and Nikki Haley as UN Ambassador, and John Kelly for Homeland Security. She’s not the only Democrat to have backed those four cabinet picks; fourteen others have done so. So the only reason for singling Feinstein out is… that she was ill for the ACA vote. That’s it. She’s one of fifteen Democrats who’ve OKed Mattis, Pompeo, Kelly and Haley, but those are votes that right now are better attacked on the (extreme lack of) merits of those particular candidates. When you can just read a list of all Feinstein’s votes without even scrolling your mouse wheel, that’s the point at which the difference between Schumer’s 83.3% score and Feinstein’s 100% score is meaningless. If you think supporting any Trump nominees is a problem, it’s a problem that a small caucus of Democrats (and Angus King) has, not a Feinstein problem. In fact there are zero Democrats or Dem-caucusing independents who’ve attempted to block all of his nominees – the lowest score being that of NY’s Senator Gillibrand, who voted in favour of Haley only.
Clearly, there won’t continue to be no data – the Senate’s work continues, and more votes will pile up and we’ll actually get a more sensible picture of how much different senators back or don’t back Trump’s policies and appointments. So how much will those percentages be worth then? More, clearly; we’ll actually have a percentage figure of how often a senator or congressperson votes with the Trump administration. How meaningful that is, though, is also debatable.
Take the following crazy proposition: Trump puts forward an extremely unqualified candidate for a cabinet post who’s likely to be very divisive, has no idea what they’re doing, and is solely there thanks to being wealthy or being a Republican bigwig. Now take a different proposition, which is probably equally crazy, and therefore apparently likely in the current climate given the first one has literally happened multiple times already; Trump at some point puts forward a significant infrastructure bill that includes actual state spending, which may need to be amended a few times, relying on bipartisan support to push it past fiscally conservative republicans. This may take, say, six votes all told including amendments etc.
So you (if you’re on the left, in the centre, or otherwise anti-Trump) might think that it would be bad to pass either of these things, you might think the Dems should back the infrastructure bill to create jobs anyway. Whatever your personal opinion, consider this; is my hypothetical infrastructure bill worth six times as much in terms of Trump support as backing, say, Betsy DeVos as education secretary? Either to you, or to Trump? This is the problem with raw vote metrics from a legislature; every vote is on at least subtly different subject matter, and not all votes can remotely be considered equal. The percentages, after a significant amount of time, may show trends to look at in more depth, but they need very careful handling, as some minor issues can require a large raw number of votes and some single-vote issues can have massive significance for a president’s authority and legislative agenda. The unpredictability of Trump and his unusual political positioning, especially on some areas of economics, makes this an even harder calculation to make in advance.
So in short, be careful when using these figures either to praise or damn lawmakers – and if you think you’re seeing a pattern, look closer than the topline data. It’s great that Nate has built this tool, but it’s a tool for enabling proper analysis not a morality score readout. It would be nice to see the option to, for example, divide votes by topic added to the system, so people can get a closer view not just of how often a legislator agrees with Donald Trump’s white house, but on a far more difficult (and probably more important) question; on what?