Does Anger Win Elections?

On Sunday, September 25, the far-right Brothers of Italy party won a commanding victory in Italy’s general election. They’ll be the leading party in the country’s right-wing governing coalition, and their leader, Giorgia Meloni, is likely to become prime minister. 

In Italy’s last general election, in 2018, the Brothers of Italy received roughly 4% of the vote. This year, they received around 26% – more than any other party in the country.  

Between 2018 and 2022, Meloni and her party rallied voters over common far-right concerns like immigration, the influence of international elites, and the rise of a supposed “LGBTQ Lobby.”

Along with ideological concerns, Meloni and her party have employed a consistent mood that’s familiar across our politics: anger. 

Today, it can seem like cultivating anger is a key to political success. But how effective is it as a political strategy? And how, exactly, is it deployed? 

These are questions that Watson Professor and Political Economist Mark Blyth, Rhodes Center Postdoctoral Fellow Nicolò Fraccaroli, and Brown University undergraduate Nadav Druker '23 are uniquely suited to answer. Using data from over 18,000 Italian Facebook posts over the last decade, they’ve analyzed the presence of political anger in the rollercoaster of contemporary Italian politics. They then devised a new and fascinating way to measure this anger, and in the process are helping shed light on how the emotion is used, and the effect it can have. 

On this episode, Dan Richards talked with Mark, Nicolò, and Nadav about Italy's elections, how the results fit into their research on political anger, and what it can teach us about politics in Italy and around the world. 


[MUSIC PLAYING] DAN RICHARDS: From the Watson Institute at Brown University, this is Trending Globally. I'm Dan Richards. This past Sunday, the far-right Brothers of Italy party won a commanding victory in Italy's general election and will be the leading party in the country's upcoming right-wing governing coalition. The party's leader, Giorgia Meloni, is likely to become prime minister.

In Italy's last general election in Twenty Eighteen, the Brothers of Italy party got about 4% of the vote. This year, as of this recording, the party is set to receive around 26% of the vote, more than any other party in the country.

Between Twenty Eighteen and today, Meloni and her party have rallied voters over common far-right concerns, like immigration, the influence of international elites, and the, quote, "LGBTQ lobby."

Along with ideological concerns, there's also a tenor a vibe that Meloni and her party have employed that's familiar across our politics today. Anger.


DAN RICHARDS: Of course, anger isn't only used by the right wing. It's found a home in movements around the world on the left and the right. It can sometimes seem like whipping up anger is the key to winning elections. But how effective is anger actually? And how exactly is it deployed?

These are questions that our guests today, Watson professor and political economist Mark Blyth, Watson postdoctoral fellow Nicolo Fraccaroli, and Brown undergraduate Nadav Druker, are uniquely suited to help us answer.

Using over 18,000 Facebook posts from political parties and their leaders over the last decade, they've analyzed the presence of political anger in the roller coaster of contemporary Italian politics. They devised a new and fascinating way to measure this anger and in the process, help shed light on how it's used and the effect it can have.

On this episode, I spoke with Mark, Nicolo, and Nadav about Italy's elections and the rise of the brothers of Italy, how the election results fit into their research on political anger, and what this all can teach us about politics in Italy, as well as, well, whatever country you're listening from.


Mark, Nicolo, Nadav, thanks so much for coming on to Trending Globally.

NADAV DRUKER: It's lovely to be here.

DAN RICHARDS: So I wonder if we could just start by recapping the election results in Italy briefly. We're recording this on Monday, so it might be a little different by the time our listeners hear this. But what are we looking at right now? Nicolo?

NICOLO FRACCAROLI: So we saw a big win for the center-right coalition. The center right, as a whole, as a coalition, got 43% of the votes. And the big winner in this coalition was the party of Fratelli d'Italia, Brothers of Italy, headed by Giorgia Mark.

And it's a big win because it used to be like the smaller party in the coalition, like the third party. And now, it became the first. And the main opponents, which was the Democratic Party, the center left, got 19% of the votes, which was around what we were expecting but a bit lower than what they were aiming for.

One final notice about the turnout, which is something that has been in a steep decline in all the Italian elections since the '80s and now, it got the lowest that it has been since then, since the postwar period.

DAN RICHARDS: So we're going to unpack some of these election results and how these parties' fates have changed in the last few years in elections but through a lens that you all have looked at in some research, looking at some of the role that anger plays in politics in Italy.

And I was wondering if we could just start with, what prompted you all to want to explore this particular emotion in politics right now?

MARK BLYTH: I wrote a book called Angrynomics.


MARK BLYTH: That might be the clue. So Angrynomics' thesis is very simple, essentially, that if you basically make millions of people, who have precarious relationships to the labor market and basically make sure that their incomes remain, if not flat, then declining in some cases in real terms, if you're in Italy, in particular, you've effectively lost 20% of GDP over the past 20 years despite running a budget surplus that's been mandated by basically your partners in the EU.

It's pretty miserable. And that tends to promote anger. And that was that. And then Nic came along and said, I wonder if we could actually measure this.

DAN RICHARDS: And so how did you go about measuring this?

NICOLO FRACCAROLI: Mark and I were doing research independently at first. And then I happened to meet Mark's book at the point in which I was actually wondering a lot about what was actually going on in party politics. Once I started reading Mark's book, I thought, oh, yeah, anger could be really an interesting trait to explore.

So Nadav and I started to look into this Facebook data that we have on political parties in Italy during the last two electoral campaigns. And we started thinking how we can extract anger based on the text of this Facebook post that parties and their leaders posted before the elections. And maybe Nadav can say a bit more about how we measure these things.

NADAV DRUKER: Right. So when you measure the effect of anger, there's two broad categories of doing that. One is looking at keywords, so that's exact search. And then one is looking at the semantic meaning of the text. And the first approach, we have two measures to do that.

One is a dictionary-based approach. One, we defined a lexicon of anger terms. And then we went and looked at each document in our corpus and measured how many words in each document is an angry word.

DAN RICHARDS: So literally, like file find, kind of just like, does it have these words?

NADAV DRUKER: Yes. I mean, there is some preprocessing. Of course, you have to stem the words just to get more capture, more of that variation. But essentially, yes, just like taking the ratio of anger terms in a document over the total number of words in that document. And that produces this anger ratiometric.

Doing similar thing with caps lock, caps lock is just sort of like a convention in social media to express angry tone. So literally, just counting the number of words in a document that are fully caps locked and then going to the semantic search results. We took two specialized BERT-based models.

So BERT is this essentially just a deep learning model that allows you to understand relationships between terms. So it works really well with sentences. You can really understand context well. So we took two models based on that. One is trained on 11 billion words, so it's pretty robust.

And then it's fine tuned with Twitter data on emotion sentiments. And then the second model that we used was the Google model that measures sentiment in comments specifically. So they crawl The New York Times comment section and Wikipedia.

NICOLO FRACCAROLI: Can I add a bit more comments on that? So the starting question was, how do we capture anger in this Facebook post? So we started really brainstorming, the three of us, about different ways to do that. And one way, as Nadav was describing, was this dictionary technique.

So you take a set of words that you think can capture anger, can be associated to anger. A simple example could be, I'm angry. So if there is the word "angry," then this is an angry post. But there are a number of other words, for example, one could be a "traitor," which is something that is appearing a lot.

And I think it's interesting because it's a word that is very recurring in the rhetoric of the far right. But this is something that could be maybe a bit too simplistic. It doesn't really capture anger as a whole. It may also not taking into account the context of an angry message.

And so that's why it was really cool that Nadav came up with this idea of looking at this machine learning algorithms that Google uses to moderate forums, like for example, The New York Times' comment section. And they have these machines that are able to identify when you're using a language that is so-called toxic. And usually, this is very correlated with anger.

So a dictionary technique is something that has been used already in academia a lot to search, for example, even the degree of populist content of a message. But what we do is to distinguish it from a populist rhetoric and just focus on a single emotion, which is anger, and then look at these other more toxic-language-based measure.

And finally, the other measure that is interesting, particularly in view of this election, is the caps lock, which is something that was a bit more creative in what we did. And it's a good measure of anger, as Nadav was describing. Indeed, people use it, so we probably use it sometimes when we want to really deliver a message that's a harsher tone than a normal message with lower letters.

It mat capture anger. It's probably strongly associated with that. It could be also not simply the anger. But I think it's an interesting measure in the way that it delivers a message, yeah.

DAN RICHARDS: Yeah, absolutely. So these measures were all working together then to get the most kind of comprehensive picture of how anger was being used. So I guess then, I have two questions. The first is, what types of parties or groups in Italy were using anger most on social media?

NICOLO FRACCAROLI: The interesting aspect of all these is that we use different measures without really knowing where they were going to lead us. And the interesting pattern-- for some, it could be surprising. For some, it may not be as surprising.

But it was that the far right or the right wing parties, in general, tended to deliver more angry messages than other parties across both the Twenty Thirteen and the Twenty Eighteen elections that were the past two elections before this one. And the party that scored the highest across all measures apart to one is the League.

So what happened between Twenty Thirteen and Twenty Eighteen is that the party was completely rebranded. It became a party that was matching more with the profile of the typical populist far right that we saw rising after the crisis. So far right that is close to the right in the US, to Trump, but also to Le Pen. And it was also more nationalist far right.

So the League in Twenty Eighteen was the party that delivered-- they supplied the most anger in this electoral campaign. And under all measures but one, it was followed by the Brothers of Italy, which at the time, was a very tiny party. It was a party that would get 4% of the votes. Today, it got 26%.

And what is interesting is that this was true for all three measures but the caps lock measure. The caps lock measure is the only one where the Brothers of Italy is actually the first party, the party that has the highest score, which means the highest usage of caps lock, followed by the League in Twenty Eighteen.

DAN RICHARDS: What do you make of that? Is it just a random style choice, or do you have any ideas about why caps lock was taken on by one party?

MARK BLYTH: Well, given the fact that Meloni and her party have gone from relative obscurity to basically being the dominant party and they were the one that used caps lock the most, at the least, you could say that caps lock is a leading indicator--


--of rising populism. But in terms of the study, overall, I mean, I think, that its real import is to show-- again, to go back to Angrynomics-- the reason people get angry is because they're frustrated. But for party politics, the important part is anger is deeply mobilizing.

I mean, if you think about Trump's rallies, what it is is basically a giant anger generator. It's a sort of a grievance fest. And that's what these parties feed upon. And regardless of what one thinks about it, it is immensely powerful as a mobilizing device. And that's what we see now.

NICOLO FRACCAROLI: I totally agree with this view. And I think it's also helpful to understand a bit better the result of the latest elections, because I think, the caps lock is just part of a wider strategy than the party has adopted, which is a strategy of catching the attention, catching the eye.

And this is something that has been implemented by many fringe parties. But the interesting thing is that this fringe party across among all the others is the one that indeed emerged as the victorious one.

But you can see it also in its broader strategy because when all the parties agreed to support the Draghi government, which is the latest government, the one that actually fell, the Brothers of Italy decided deliberately to be the only large party and the opposition.

And that really rewarded them because all the voters from the League decided to shift from the League to the Brothers of Italy. I think this is telling us something because it tells us that this strategy works. But it's also telling us that we shouldn't be worried about Italy shifting ideologically to fascism or to the-- or to this very extreme far right.

It was far right already before because these votes were votes that would go to the League, to an anti-migrant party. But they simply shifted from one party to the other because this party was proposing the alternative, whereas all the others were kind of part of the same cluster and that the same sort of technocratic but with political government.

MARK BLYTH: Yeah. There's a weird thing about it. Before, we had the term "populist." We used to call them things like insurgent parties, right? So there was this idea of all the mainstream parties were like together in a club or a cartel, if you will.

And the way that you bust up cartels is you get these upstart interests that come in and disrupt things. Well, what happens with most of them-- and Austria was a great case of this-- is that they had one of these parties 20 years ago. And what they ended up doing was joining the coalition.

It was just kind of like whining to get your share at the table. And if you mobilize this kind of outsider anger thing but then become part of the mainstream, then your position gets really fragile. And I think that's exactly what we've just seen.

NADAV DRUKER: I just want to add a small caveat in the measures. So I think what was surprising to us in the process was to discover that it's the same player scoring angriest like repeatedly on each measure, because we expected some variation. And it was really just surprising to see that the right is scoring so high by a lot, sometimes by like double that of any left party.

DAN RICHARDS: The law has been made lately of how Georgia Meloni has also started to take a more moderate tone rhetorically. And I wonder, what do you think that shows about this role of using anger and the limits of it, maybe? Or just what do you make of that change?

NICOLO FRACCAROLI: This is a very complicated question. I don't feel like I have a proper answer to that. But what really brought to my memory is the election in France and the run off between Le Pen and Macron.

And I remember Le Pen trying to take more moderate positions as well when it came to indeed convincing voters that are less extreme, that are more moderate, that are more in the center to support her.

I think Meloni knew who her opponents were. And I think she knew that the risk would be to be labeled as the irresponsible leader that would have put Italy in a very difficult geopolitical situation.

MARK BLYTH: There's an American example for this. You may remember Mitt Romney. So Mitt has made this lots and lots of rather incendiary claims during the primary. And then his campaign manager famously said that, well, that's all done now. Now we'll have a reset.

DAN RICHARDS: Very shake the etch-a-sketch maybe.


MARK BLYTH: Exactly. So in other words, that's quite clearly, we are mobilizing people with anger. But now, we've mobilized them, and I've won a nomination. I will stop doing that, and I will try something else. Now, whether you can get away with that, once you've publicly said it, I mean, that actually really hurt Romney's campaign.

And the counter case again is Trump. I don't think he thinks of it this way. But what is effectively doing is he really cares about the marginal angry voter. He wants to get just enough to get elected. He has no intention of converting anyone else on the other side, however that is defined.

And that's a much more, if you will, coherent anger management strategy, if you want to put it that way.

DAN RICHARDS: Anger management strategy, I like that. Do you guys have any plans or desires to try and use these types of measures in other elections or other types of movements around the world, because this is clearly like a global trend?

NICOLO FRACCAROLI: What we are actually trying to do now is to dig deeper into these results. So this is-- it's important to say it's still a work in progress. So I think, now, we feel like fairly good about the strength of these measures. But what we want to understand is, what could be the actual implications of these measures, or where is the anger coming from?

So it was more difficult to distinguish as Mark did in his book with Eric, between anger as a moral outrage or anger as a tribal--

MARK BLYTH: Identity.

NICOLO FRACCAROLI: --identity, yeah. But we can see that most of the anger that we would identify would be more this negative side of anger, because let's remember that anger is also-- it can also be positive. It implies that the population is dissatisfied with something.

And usually, the definition of anger implies that there should be a payoff. So indeed, it requires the politicians to act and deliver something. But this side of negative anger is indeed very close to a sort of identification into a tribe. So we see that there is a lot of that.

And a lot of this anger is also addressed toward mainly political opponents but also migrants, in the case of the far right and very rarely, to other actors such as journalists. So there are some parts from the five-star movement that is also very critical to journalism because it's often perceived as sort of the idea of the establishment.

So what we want to do with this data is to understand why the angry parties, for example, are more successful in more economically distressed areas. And then we want to understand really whether anger itself has some impact, also especially in terms of social engagement-- so the number of likes that these posts get compared to others.

MARK BLYTH: And one other thing to consider as well is it's just harder to be perpetually angry when you're on the left because imagine your message is inclusion. Your message is tolerance. It's very hard to get angry about tolerance and maintain any straight face or credibility.

So there's also something in just the nature of the messages themselves, whether they can become anger vectors. And maybe, there's a built-in structural advantage for the right on that one, in the sense that it's about you've defected from the club, you're a traitor.

These people are a threat, et cetera, which is very different from, we need to build a better society. It's hard to imagine someone shouting that one out.

DAN RICHARDS: Right. Well, and that makes me think of it, at least in the US, some of the most successful far-left or just on the left movements have employed anger to the extent they can of us versus them, 99% versus the 1%. In fact--

MARK BLYTH: Sunrise is another example. Absolutely.


MARK BLYTH: But I mean-- but then also, the lesson is in terms of the overall balance of forces, how successful have they been? Not even in their own party.

DAN RICHARDS: What you were just saying, Nicolo and Mark, brings me to my last question, I think, which is, how is this research changed or made you all reevaluate some of how you view anger in politics?

NADAV DRUKER: Well, I have to say it's not making me really optimistic because I have to seek the ulterior motives in any message that I read on the news. But I think that just us talking about it right now and doing research about it is, for me, at least kind of goes to show that we're aware of this trickery or this ploy of mobilizing us via anger.

NICOLO FRACCAROLI: So my side, I think this research was part of a sort of longer personal path as well. I got interested in the topic mostly during the pandemic because I was feeling-- I was also feeling angry. And I would see many people around me feeling angry at the government and the government not taking action or not taking action as we would have liked.

I could notice how then angry messages delivered by different politicians would have some sort of appeal. They wouldn't go as far as shifting my voting preferences, but I could definitely see how they would act in that way. And I think with this study and the research that surrounded it, it allowed me to see anger a bit more from its multifaceted perspective.

So indeed, understanding how there could be a negative side of it, which is really like just screaming or writing in caps lock and vis-a-vis the anger that would be a more constructive anger-- so the one that Mark and Eric defined as moral outrage and which is probably the anger that we need in many parties that give the wrong impression, that give the impression that there isn't really much to be changed.

And these are the parties that are usually not considered as populist. But they are considered as the establishment. And so what I got from this is that we would need a bit more emotions also from that side. And emotions are not necessarily something completely irrational.

We tend to have this distinction between rationality and emotion. But sometimes, I think, the two things may not be as separated as they look like.

DAN RICHARDS: Nadav, Nicolo, Mark, thank you so much for coming on to talk with us.

NADAV DRUKER: Thank you for having us.


DAN RICHARDS: This episode of Trending Globally was produced by me, Dan Richards and Sam McKeever Holtzman. Our theme music is by Henry Bloomfield. Additional music by the Blue Dot Sessions.

If you like this episode, be sure to check out our other two podcasts from the Watson Institute that feature Mark Blyth as host. One is called Mark and Carrie, and the other is called the Rhodes Center Podcast. We'll put links to both in the show notes.

And be sure to subscribe to our show if you haven't already wherever you listen to podcasts. If you have any questions, comments, or ideas for episode guests or topics, shoot us an email at trendingglobally@brown.edu. Again, that's all just one word, trendingglobally@brown.edu. We'll be back in two weeks with another episode. Thanks.

About the Podcast

Show artwork for Trending Globally: Politics and Policy
Trending Globally: Politics and Policy
The Watson Institute for International and Public Affairs

About your host

Profile picture for Dan Richards

Dan Richards

Host and Senior Producer, Trending Globally