Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

What is fascism?

July 13th, 2016, 8:31am by Sam Wang

We have a few weeks of waiting to see where polls are headed. I expect many undeclared voters to choose up sides after the conventions. In the meantime, the political theater is a more riveting story. Not necessarily in a good way.

For those who take the long view, Umberto Eco’s essay defining “Ur-Fascism” is a useful look back. I learned of it from Jamelle Bouie’s piece. The point here is to get away from “fascism” as a cheap insult, and ask what were the common factors in the mid-20th-century in what one could call classical fascism.

Eco lists 14 criteria that may seem familiar to observers of the current scene. The criteria are interesting because they were laid out by him long ago, independently of any current events. It could be turned into a checklist, like the DSM-V has for mental illness: meet 5 or more criteria, and it’s time to get evaluated by a professional.

I see seven items that potentially match current events:

  • a cult of tradition
  • rejection of modernism
  • opposition to analytical criticism
  • appeal to a frustrated middle class
  • obsession with a plot
  • humiliation by the enemy
  • machismo

There may be more – see the list for yourself.

Eco closes with a quote by Franklin D. Roosevelt from November 4, 1938: “I venture the challenging statement that if American democracy ceases to move forward as a living force, seeking day and night by peaceful means to better the lot of our citizens, fascism will grow in strength in our land.”


Okay, back to polls. Yesterday the electoral vote estimator and the Presidential Meta-Margin finally moved – sharply, by 0.8% toward Donald Trump. This was caused by a string of four Florida polls favorable to him: two partisan pollsters plus Gravis and Quinnipiac. If this shift is real, it looks like it happened within a few days of June 25th. What would be the cause of that? Waiting to see whether it’s lasting.

Today’s final installment of “waiting to see”: the generic Congressional preference is staying high, at Democrats +7.5%, while President Obama’s approval is back to its plateau of Approve +1.5%. Under the view that the November election is a referendum on the incumbent President, these numbers and the Clinton v. Trump race should converge a bit in the coming months.

Tags: 2016 Election · House · President

64 Comments so far ↓

  • whirlaway

    Not just Florida polls. RCP shows that there is a swing towards Trump in OH and PA as well. If Trump wins those three states and manages to hang on to the Romney 2012 states, he is the next POTUS…

    • Matt McIrvin

      He has to get ’em all, though–no margin for error unless he can also win some bluer states. Currently his campaign is dumping money into ridiculous things like trying to flip New York. They have a lot of wising up to do in a short time.

  • Tony

    The polls yesterday were weird. The national polls were same as usual, showing Clinton maintaining her normal lead. But the state polls, other than the one in Colorado, were bad for Clinton. Even if you ignore the two republican leaning polls, that’s still 4 state polls (2 Florida, Iowa, and Pennsylvania) showing her down in swing states, and one showing her tied in Ohio.

    One would think either the national polls or state polls are of, and if the shift is real it’s due to her emails.

    • Commentor

      Much has been written about Trump’s weakness in certain red states, which can explain weak national numbers even when he’s comparatively strong in some states.

    • Matt McIrvin

      Clinton is a bit down in the national polls too–new polls from NBC/SurveyMonkey and McClatchy/Marist are down.

      I don’t think it is too surprising. I would not be at all surprised if there was a real, small shift toward Trump. The email business was temporarily in the news again with the end of the FBI investigation and Comey’s public statement, which were not flattering to Clinton though they removed the possibility of an indictment. (I would expect the salience of that for most people to fade rapidly.)

      And then we had the horrible triple hit of two high-profile police shootings of black men and the Dallas massacre, which brought racial conflict front of mind–Trump behaved awfully through the whole thing, which limits the gains he can get, but it probably helps him on balance when people are generally scared and angry about race and violence.

      Convention time is coming up early this year and I would expect some wild gyrations in the polls. I wouldn’t be too surprised to see Trump ahead nationally for a week or two, even though a lot of the Republicans are treating their convention like a potential dumpster fire.

      But the Democratic convention comes second this time, and aside from some increasingly irrelevant Bernie-or-Busters around the margins it looks like it’s going to be a much friendlier event.

  • Adam

    A little unnerving to see such polls, but I feel better that a) they are all from one polling place and b) are still in the wonky summer polling period.

    I am suspicious of any poll showing a huge swing of non-whites towards Trump, however. In short, Clinton may find her path easier in Florida vs an Ohio this year.

    We will likely see more polls head towards Trump in the next two weeks with the convention bounce.

  • Matt McIrvin

    Gravis and Quinnipiac have tended to give low-ish numbers for Clinton’s national support all through the season, though it’s anyone’s guess whether that makes them less accurate. Quinnipiac’s last Florida poll was Clinton +8, though–a lot of variance there. And it looks as if Gravis hasn’t been polling Florida until now.

  • Olav Grinde

    Historically, how strong is correlation between the generic Congressional preference and the popular vote results in Congressional elections?

  • Eric

    The Florida polling link in the article still shows Clinton up by 2.2% on average in Florida.

    • Sam Wang

      Yes, that is what HuffPollster’s smoothing algorithm reports.

    • Matt McIrvin

      What models say about Florida at this point is going to depend heavily on their averaging algorithm, because it’s a huge swing, but on the basis of a cluster of recent polls that are either not-well regarded or have a Republican house effect. Sam’s model tends to try to deal with crazy outliers by using medians, but the median here genuinely moved.

      538 has Florida still blue because they either down-rate those particular polls, adjust them to compensate for past house effect, or both. Whether that’s legit is open to question. If you do it in a motivated way you get travesties like the Unskewed Polls site in ’12. Probably better to do something simple, wait and see–there may have been a real shift here.

    • Michael Hahn

      I just got back from vacation, and wondered what happened to swing Florida over to Trump. After reading here, it seems that it may be an artefact of a concentration of “unique” polls. But still it sent a bit of a shiver down my spine to think that Trump might have a chance of making it close. Sam, do you think this sudden shift in Florida is an anomaly, or is there any indication that there has been a substantive shift in Florida that crept up on us unnoticed?

  • bks

    I think this is on topic, though others may differ:

    Many Voters Think They’ve Seen Trump Ads On TV — But He Hasn’t Run Any

    • Josh

      It’s interesting–my understanding is that running ads in general has very little effect overall on a race. Clearly there are some exceptions like Kerry getting swiftboated and Johnson’s nuclear bomb ad, but for the most part, it seems like all but the most successful ads don’t accomplish much. Maybe Trump’s campaign is hip to this?

    • alurin

      @Josh; though usually there are ads by both sides. this election will be a test case of what happens when one side massively out-advertises the other.

    • MH

      I am very curious about what the rates of “regular” TV watching are for younger people. I’m not super-young (35) but I have not had cable for years and only stream TV shows and movies via services like Netflix and HBO Now. As a result, I never see political TV ads. I thought I was an outlier, but I wonder if the under-30s out there are in a similar boat. It seems to me that social media presence is what matters more, and Trump’s Twitter feed gets an insane amount of attention online.

  • E L

    Thanks, Sam. Trump certainly has fascist tendencies but I don’t think he is a full blown fascist who openly advocates street violence against opponents… yet. He’s still more Berlusconi than Mussolini; however, he could shift if he thinks outright fascism with street violence is a winning position.

    • Commentor

      1) Did any historical fascits immediately advocate street violence against opponents before gaining first positions of power?

      2) Is it really true that Trump hasn’t advocated violence against opponents? Hasn’t he suggested that his supporters at rallies should punch opponents and/or that he’d pay their legal bills if they did?

    • Olav Grinde


      #2 is on a long list of things that Trump has said, and later claimed not to have said. Just saying… ;)

    • pechmerle

      Sure, there were historical fascists who advocated street violence prior to gaining first positions of power. This is true of both Hitler’s Nazis and Mussolini’s Fascists. Google Hitler and Beer Hall Putsch, and Mussolini and March on Rome.

  • Mark F.

    Trump needs OH, PA and FL. If he loses any of those 3, Hillary will win. Skeptical he can do it, especially Florida.

  • Gopalan

    Trump does better when he is quiet for a few weeks (relatively speaking). But the big elephant in the room is the undecideds, which as Sam pointed out is relatively large this year (15%?)

    Brexit metamargin was 0.8+/-1.x% on the eve of the vote, but 9% undecided (by PEC numbers). Could a large swing in undecideds here still overwhelm the +2-+3 % metamargin for HRC?

    • Brian

      My understanding is that one of the main takeaways from previous cycles’ meta analysis is that “the undecideds” are mostly a myth. Generally speaking, the relatively small percentage of people who self-identify as undecided or independent tend to either evenly distribute, or are rendered irrelevant because pollsters’ models bake-in their effect. I could be wrong, but this is what I inferred from the infamous 2004 “undecideds will break late for Kerry” mistake on this site.

    • Sam Wang

      Yes, but that is only the case in the home stretch. Current undecideds are a different story.

    • Olav Grinde

      Has any comparison been done between how large the undecided group is in two-candidate polls and three- or four-candidate polls?

      Is it true that in many Trump–Clinton only polls, responders that have decided to vote for Johnson or Stein, may incorrectly be pigeonholed as undecided voters?

      If there is a significant difference, then that would seem to spotlight faulty polling methodology.

    • Suvro

      I wonder if the larger proportion of undecideds, as of a few days ago, is because:

      a) some Republican leaning voters are still waiting to see what will emerge from the convention, and likely they are not strong supporters of Donald Trump, and

      b) some Sanders supporters were waiting to see how strong his support for HRC would be.

      If these two assumptions are correct, the fraction of undecideds should go down by early August once the two conventions are over.

  • Mark R

    Nate Cohn at the NYT has some color on one of the “partisan” FL polls, JMC: “Both are automated surveys that have no means to contact voters who don’t have landline telephones. It’s no surprise that both show Trump ahead among Hispanic voters . . . The JMC Analytics poll doesn’t appear to be weighted by race, and it has an electorate that has no resemblance to the actual Florida electorate. I suspect it’s even worse on age, but it doesn’t even mention that.”

  • anonymous

    I know that talking about individual polls and their methods is verboten here, but since there was a significant move in the meta-margin, a contrary perspective on the new swing state polls (by a partisan group) can be seen at:

  • donnie

    Eh, I don’t know Sam. I went through the trouble of looking up what fascism was way back when Trump was first being labelled as such. As near as I can tell no one agrees on what fascism is. In most cases, it’s just used as a scare word to mean whatever the speaker wants it to mean.

    I appreciate Umberto Eco’s 14-element definition of fascism, but honestly if a word has a fourteen-element definition with a “pick any” quality to it, it’s not really a word. It’s a list. So I think it’s fair to say that “fascism” is – first and foremost – not an actual word with any agreed upon meaning in the English language. Which is probably why when most people use it they are literally talking nonsense.

    Maybe fascism had an agreed upon meaning once upon a time when it was a real tangible, threat to Western society. But as of today fascism has devolved into the sort of insult you use when you have bad feeling about a person in power (or potential person in power). 9 times out of 10, accusing someone of being a fascist in today’s world is simply an abandonment of reason.

    • Sam Wang

      I agree that as an insult, the word is not helpful. The point here is that there is a classical definition that one could use to gain a greater understanding of the current political scene.

      I am not deterred by the large number of points, since they were established beforehand and therefore cannot be altered to suit a new set of circumstances. I encourage people to read this essay.

    • SRS

      Long lists with a “‘pick any’ quality” are a necessity for proper scientific cultural analysis.

      Cultural and ideological movements like Fascism aren’t like chemical elements that you can extract from their native environments, purify, and then examine in a test tube.

      They can only exist in the wild and in the wild these movements always adapt to their current environment.

      Fascism in Germany was not the same as Fascism in Italy which was not the same as Fascism in Spain which was not the same as Fascism in the UK etc. etc.

      Ultimately, all human beings are psychologically unique and thus all societal movements are unique to their specific time and place. Calling something ‘Fascist’ is never an objective discovery of a factual true nature innate to the person or movement, it is an analytical distinction that can be invoked if it helps a researcher establish cross-cultural patterns.

      Each Fascist movement is unique, and each can be understood and researched as its own unique local phenomenon. But looked at across cultures Fascist movements also draw on certain common psychological and ideological underpinnings.

      A proper qualitative understanding should draw on both. It should understand, in this case, Trump as both a local uniquely American phenomenon *and* as the American expression of a broader phenomenon that has counterparts across the globe and world history itself.

    • donnie


      I think you just proved my point. If fascism doesn’t mean the same today as it meant 80 years ago, and if it doesn’t mean the same thing in one country as it means in another, then I hardly see how the word can have any meaning at all.


      I agree that if we’re going to talk about the current political scene in terms of “fascism” then we need an agreed upon definition. And if Eco’s 14-point checklist is the best one we have I’m willing to accept that. It’s certainly better than having a nonsense word that’s just thrown around irresponsibly to scare people.

      But in that case, if anyone wants to assert that Trump or some other person is a fascist, I’m going to have to ask them to explain that opinion based on the 14-point elements laid out by Umberto Eco. And quite frankly I don’t see how Eco’s checklist works for any modern political figure that I’m aware of.

    • Sam Wang

      The reason the checklist appealed to me is that I was able to find seven points in the modern situation without difficulty (cult of tradition, rejection of modernism, anti-analytical criticism, appeal to a frustrated middle class, obsession with a plot, humiliation by the enemy, machismo). It’s like a DSM-V checklist: achieve some minimum number and further questioning is merited.

    • Michael Coppola

      There are an awful lot of words that don’t mean the same thing today as they meant 80 years ago, and don’t mean the same thing in one country as they mean in another. It’s a wonder we can converse at all.

    • 538 Refugee

      Language is a living thing. Language was my lowest ACT score. (Not sure what it is like now but it was multiple sections when I took it ‘back in the day’.) It was the exact median score. I always liked to think of that as a perfect score misinterpreted. After all if I spoke English like almost everyone else. Wasn’t that ideal? ;)

      Even if Trump is a fascist we have those pesky checks and balances in place.

  • Alex

    It shouldn’t be surprising that there’s a Comey effect– she had a brutal week. I suspect she’ll get a small Bernie bump as he begins campaigning with her.

    • Olav Grinde

      The Bernie endorsement could have been better-timed. It seemed to drown in other news, mostly negative for Hillary.

    • Matt McIrvin

      I’m not sure the Sanders endorsement was ever going to have any immediate short-term effect–the Bernie-or-Busters just came up with reasons he didn’t really mean it, or condemned him as a sellout, and moved on to advocating Jill Stein or Gary Johnson.

  • MNP

    Interesting. But I tend to give the most credit to defining fascism to Il Duce

  • Olav Grinde

    Rather than massive TV advertising, I think it would be far more cost-effective to establish multiple companies that carry out biased presidential polls – and to carefully time their release. Deservedly or not, polls that claim “momentum” or big swings in voter sentiment get a lot of media attention.

    I think this is a topic that deserves some attention: polling that aims to create a specific narrative, rather than accurately measure voter sentiment and intention.

    Sam, to what extent do you feel reason to believe this is taking place?

  • gorram

    Arguably, fascism is so hard to define because it’s an ideology built around misdirection. Laclau and some other theorists (directly and indirectly) responding to Eco made basically that point a while ago:

  • Mark F.

    Is he really trying hard to win New York? Trump may say he is, but I doubt he’s actually that stupid. Obviously, he needs to concentrate hard on FL, PA and OH plus the usual swing states.

    • donnie

      Seems to me his focus on winning NY is a classic “thinking past the sale” Trump gambit. Pundits will sit around and debate for a whole news cycle whether he can realistically win NY, instead of sitting around and debating whether Trump can realistically win the election.

      I’d be shocked if he was actually putting any of his own money into such an endeavor.

  • Mark F.

    I wonder how likely it is that Trump can pull a George Bush, i.e. lose the popular vote but win the election? These polls suggest that as a real possibility again.

    • Matt McIrvin

      If there’s a gap of a few percent between Clinton’s national lead and the Meta-Margin, that’s the opening for a Trump win with a popular-vote loss. 538 rates it as only a few percent probability, but that’s partly because their probabilities for everything are really fuzzed out, so the part of the distribution in that window is smaller.

  • Mark F.

    This is Trump’s winning electoral map IMHO:

    • bks

      I think he gets Iowa in that scenario. If it’s close, the Keystone State will live up to its nickname.

    • Jay Sheckley

      Mark F: Your made my heart stop. But I couldn’t make a map matching yours without turning every state pinked by aggregated state polls red, then showing a major proTrump shift among states not the darkest blue. I don’t see how these blue states were selected to become red, unless it was for their high EV. bks’ point seems valid: Why not Iowa? If Pennsylvania is plausibly red, why not Virginia, New Hampshire and Nevada? If this map is just to show that Trump has a chance, the header here gives his election a 20-35% chance. From a fifth to over a third is a real chance, neither favorable nor inconsiderable.

  • Doug Kiel

    As a 60 year old political scientist, I share Sam’s concern with “fascism.” Perhaps a better term, but also a bit fuzzy, is “authoritarianism.” Marc Hetherington’s (Vanderbilt Univ.) work shows that it is relative levels of authoritarianism that really divide the political landscape. In short, basic psychological constitutions (open vs. closed minds) are more important than policy or ideology. History suggests that human (political) institutions are more fragile than we might like to think. We should be worried.

    • bks

      Authoritarian Anarchism?

    • Jay Sheckley

      Yes authoritarian is the better word for people’s resonance…with fascism. On Facebook a few months ago, blonde bombshell Mamie Van Doren pointed out that in the 1990’s Trump’s then-wife Ivana said in an interview re:DT’s political leanings that he kept a copy of Mussolini’s speeches on his night stand. Though the candidate proudly retweets Mussolini quotes, when I researched this, the bedside book is apparently Mein Kampf. Go figure. Though my data trail seems a bit fuzzy, surely “Sam’s [shared] concern with fascism” is accurate. Here’s a link to a newer article about this, headlined “7 TAKEAWAYS FROM VANITY FAIR’S 1990 PROFILE OF DONALD TRUMP”

  • Just Dropping By

    As someone who wrote a final paper on fascism for a political ideologies and philosophies class in undergrad, I salute you for looking to “to get away from ‘fascism’ as a cheap insult….” But I would venture to say that if you’re looking to describe the likely form of governance under a hypothetical Trump administration, “sultanism” is a far, far better fit based on the evidence available so far:

    • Olav Grinde

      Sultanism ≈ Narcissistic authoritarianism

    • donnie

      I am naturally skeptical of all of these “President Trump will be some kind of crazy, never before seen radical authoritarian” speculations. Last I checked the president has very little power in the grand scheme of things.

      That being said I find your contention that a President Trump would best described as a “sultanist” to be extremely amusing.

    • Jay Sheckley

      Thanks, Just Dropping By. You’ve helped make sense of the senseless:
      “The sultan … is never bound by any rules or given ideology, even his own.”

    • Doctor Science

      Thank you, JDB! You’re right, Sultanism is a close fit, very accurate. And it even ties in Trump’s taste in decor …

  • Lorem

    I don’t think it’s useful to think about whether someone is fascist or not. I think a better consideration is something like “to what extent do each of these political positions predict authoritarianism (or whatever else we dislike) when in power”, which is similar, but a bit more focused.

  • Phil

    Worth noting that similar conditions existed in the US at the same time, Our solution was FDR’s Democratic Socialism.

  • Robert Johnson

    I am not a political scientist, but I think the term Fascist is a lot easier to define and less problematic than most people make out. In fact, I would say that politicians and political scientists and pundits are largely to blame for the fuzziness about the term. It has been used too many times to refer to an ambiguous disliked ideology.

    There is a very simple way to identify fascism. Read Mussolini.

    To me, Mussolini defined Fascism sufficiently well that we don’t need to argue over the term.

    If you have some ideology that you don’t like, but it doesn’t fit Mussolini’s definition of Fascism then it is not Fascism and some other word should be used.

    Dr Wang, whether you agree or not, your math background should at least make you sympethetic to my desire for a fixed, crisp definition.

    Absolute authority of the state.
    Organization of society to serve the state.
    Combination of state and industry for the greater strength of the state.
    Subservience of the good of the individual to the good of the state.
    Glorification of death in military service to the state.
    Rejection of peace in favor of wars to expand the domain of the state.
    Rejection of democracy in favor of a state mandated order and a state mandated morality.

    It has been a while since I read Mussolini, but that is my recollection.

    Since the term came from the Fascists of Mussolini’s era, and was based on the roman fasces (sticks bundled together for strength, sometimes with an axe head) as a symbol of strength, I think they can define the term and modern political scientists can use it or make new terms that more precisely define what modern political scientists are talking about.

  • Edge Oforever

    despising intellectuals, science
    criminalizing political opposition
    cult of personality
    state control of media/information
    merging of state/corporations
    identified scape goat(s)
    aggressive, militaristic stands

    • Suvro

      There is a very interesting read by Adam Gopnik on New Yorker titled “Being Honest About Trump”.

      The relevant passage:
      As I have written before, to call him a fascist of some variety is simply to use a historical label that fits. The arguments about whether he meets every point in some static fascism matrix show a misunderstanding of what that ideology involves. It is the essence of fascism to have no single fixed form—an attenuated form of nationalism in its basic nature, it naturally takes on the colors and practices of each nation it infects. In Italy, it is bombastic and neoclassical in form; in Spain, Catholic and religious; in Germany, violent and romantic. It took forms still crazier and more feverishly sinister, if one can imagine, in Romania, whereas under Oswald Mosley, in England, its manner was predictably paternalistic and aristocratic. It is no surprise that the American face of fascism would take on the forms of celebrity television and the casino greeter’s come-on, since that is as much our symbolic scene as nostalgic re-creations of Roman splendors once were Italy’s.

Leave a Reply to anonymous (Cancel)