Should Democrats Move Left in Down-Ballot Races?
“Bernie would have won.”
If you’ve been following progressive politics (or just browsing random corners of the internet) over the past several months, you’ve probably seen this argument. The basic idea is if Democrats had nominated Bernie Sanders in 2016, they would have kept working-class whites in the fold, prevented Donald Trump from winning key states like Pennsylvania, Wisconsin and Michigan, and kept him out of the White House.
It’s impossible to know if that argument is true -- hypotheticals like that are inherently unprovable. But some progressives have been making an interesting move of late: They’ve been importing the Bernie-would-have-won argument into discussions about how Democrats can win down-ballot going forward.
Specifically, they argue that if Democrats were to pull to the left on economics and emphasize those issues, they would be able to win majorities in the House, Senate and a number of states. Some have gone so far as to say that this is the best way for the party to make a comeback in swing or right-leaning regions. And the energy on the progressive left makes it worth asking: Would Bernie-izing the party solve its electoral problems?
I looked at past data and found that while such a strategy might not help Democrats win back the House and Senate, it could help them win back the presidency.
Why Bernie-fication Might Not Win Back the House
There are a lot of ways to assess how much ideology affects election results, but I took a relatively simple quantitative approach. I tested whether more liberal Democratic House candidates tended to perform better than more conservative Democratic House candidates after controlling for the partisan makeup of each district. In other words, if pulling to the left helps Democrats win House seats, it should show up in the data.
First, I tested whether members of Congress with a more liberal DW-NOMINATE score tended to get more votes after controlling for the presidential vote in their district. DW-NOMINATE is a measure of ideology based on how often a legislators vote together. So if a Democrat almost never votes with the Republicans (e.g. Massachusetts Sen. Elizabeth Warren), she will have a much more liberal DW-NOMINATE score than a more moderate Democrat who crosses the aisle more frequently (e.g. West Virginia Sen. Joe Manchin). DW-NOMINATE has a number of important limitations (it can only be calculated for incumbents; it doesn’t weight the most important votes; it’s divided into two dimensions instead of one, three or many of the other defensible choices, etc.) but it generally does a good job of capturing whether a congressperson is liberal or conservative on economic issues.
And from 2004 to 2014, economic liberalism did not help incumbent Democrats hold their seats. In some years, after accounting for partisanship, DW-NOMINATE didn’t have a statistically significant affect on election results. In other years a more conservative DW-NOMINATE score helped Democrats win more votes. This probably reflects the ability of more moderate and blue dog Democrats to hold districts that Republicans such as George W. Bush won.
Obviously, DW-NOMINATE has its faults, so I ran the same tests using an alternative measure known as CFscores. The idea is simple. If someone donates to two candidates, we would expect those candidates to be ideologically similar -- and CFscores use that logic along with vast data sets on political donations to estimate the ideology of political candidates on a left-right scale. CFscores also have their own advantages (e.g. can be calculated for non-incumbents) and disadvantages (which I’ve written about), so any conclusions should be taken with a grain of salt.
That being said, CFscores didn’t show an advantage for more liberal House candidates either. I sliced up the data multiple ways and added variables like incumbency but didn’t find much evidence that greater liberalism would help Democrats win swing and red districts.
There are inherent shortcomings to this type of analysis. There are reasonable criticisms of DW-NOMINATE and CFscores, and neither score purports to fully encapsulate something as complex and multifaceted as ideology. This regression also doesn’t account for every important aspect of a congressional race, and someone might come up with a different model that disagrees with mine.
It also fails to account for the broader conditions that might emerge in 2018. If President Trump has an especially low approval rating or if there’s an economic downturn or unpopular war, many voters will want to punish the GOP. In that case, some would likely cast their vote for a Democratic candidate regardless of whether that candidate would be too liberal for them under normal circumstances.
Perhaps most importantly, some might argue that a more liberal DW-NOMINATE or CFscore doesn’t correspond to what makes Sanders distinctive. Sanders supporters don’t just like his policies -- many appreciate what they see as authenticity, consistency and an ability to connect with voters. Maybe Sanders’ persona, not just his stance on issues, helped him in the 2016 Democratic primary. And while those are all potentially valid points, they’re an argument for a separate analysis (since we’re dealing with ideology here).
But it’s worth noting that this quantitative approach jibes with the recent history of the House. Modern Democrats haven’t been able to build a majority with only highly progressive members. In 2006 and 2008 Democrats still had a significant number of fiscally conservative and/or socially conservative members from right-leaning areas. Before that, Democrats were still able to win some non-minority-majority Southern seats. These Democrats weren’t really Sanders types -- the Vermont senator is liberal on a large number of social and fiscal issues. And if Democrats want to maximize their chances of winning back the House, they’ll have to find candidates who can win in normally inhospitable areas.
Sanders-ization Might Not Be Enough in the Senate Either
A quick examination of recent Senate results shows that Sanders-ization might not help Democrats take back the upper chamber.
Democrats have multiple paths to winning back the Senate. But many of those paths run through red America. Democrats have a strong hold on a number of the nation’s most populous, racially diverse and urbanized states (e.g. California, New York, Illinois), but each of those states only has two senators -- the same number that many thinly populated, rural, often red states have. So it’s worth asking whether Bernie-style Democrats are able to win and hold the swing or red states they need to reclaim the Senate.
Vox’s Andrew Prokop has pointed out that in 2016, moderate Democratic Senate candidates, not full-throated progressives, tended to outperform Hillary Clinton in red states. Missouri Democrat Jason Kander, who was more famous for his cultural conservatism and anti-corruption policies than any Sanders-esque plank of his platform, outperformed Clinton by double digits. Indiana’s Evan Bayh also outperformed Clinton, but that was at least in part due to his strong personal brand in the state.
And progressive candidates in Sanders’ mold don’t always perform especially well. For example, Ohio Sen. Sherrod Brown outperformed President Obama by only three points in 2012 -- which is no better than what we might expect from a generic Democratic incumbent senator in that state and year. Prokop also points out that Sanders-style progressive Russ Feingold ran a couple of points behind Hillary Clinton in Wisconsin in 2016. That’s about what we should expect from a generic Democrat running against an incumbent (Sen. Ron Johnson).
Moreover, Democrats who have successfully won in blue states have often done so by taking conservative positions on key issues. West Virginia’s Joe Manchin is pro-coal. Pennsylvania’s Bob Casey Jr. is pro-life. Montana’s Jon Tester is pro-gun. North Dakota’s Heidi Heitkamp is far from liberal on all issues. We could go through an exhaustive list of every Democratic senator from a swing or red state, but the point is clear. There are some liberal senators from swing and red areas (e.g. Wisconsin Sen. Tammy Baldwin) but successful Democrats outside blue America often win by compromising and moderating rather than veering sharply to the left.
But It Might Work on the Presidential Level
While a Bernie-ized Democratic Party might not be well-positioned to take back either chamber of Congress, it might be able to take back the White House. You could make this case in a number of different ways, but the simplest one is to look at the 2016 election results. A Bernie-esque Democratic Party might be able to win the presidency because the current party already got close to doing so in 2016.
Hillary Clinton won the popular vote, and Trump’s margin was narrow in the states that put him beyond 270 electoral votes. If one substantive detail of the campaign had changed (e.g. the timing of the James Comey letter and the “Access Hollywood” recording were flipped) then the election might have turned out differently. So if a more Sanders-esque Democratic Party could do everything the current party did while adding one piece to the coalition or winning back a couple of key states, it could take the presidency.
Ideology isn’t Everything
Finally, it’s worth noting once more that candidates aren’t the only important moving part in politics. Conditions matter too. If Trump has a terrible approval rating in the fall of 2018 or 2020 or if there’s a significant economic downturn, Sanders-esque Democrats might be able to win in some places where they’re not great ideological fits. Candidates can move the needle, but broader conditions often set the stage and limit the possible electoral outcomes.
So Democrats could nominate a number of very liberal candidates in less than liberal places and manage to win in a wave anyway. But if they want to increase their chances at winning the House and Senate under any conditions, they might want to pick and choose where they nominate the most progressive candidates.