When Did Anger Become a Business Model?

I happened to walk past a television playing Fox News the other day, and something stopped me. I wasn't really listening to the story. I was listening to the emotion.

Anger. Fear. Outrage. Another enemy. Another reason to be furious.

And I found myself wondering: Is this simply what partisan news sounds like — or is anger itself part of the product?

That question bothered me, because the more I thought about it, the less this seemed to be a story about Fox News.

Rush Limbaugh was doing something powerful with anger long before social media existed. Cable news discovered that political conflict could hold an audience. Then Facebook, YouTube, Twitter — now X — TikTok and an entire generation of political influencers arrived with something television and radio never really had: They could measure our anger in real time.

Every click. Every comment. Every share. Every furious reply. And every additional second we stayed on the screen.

Which made me realize I had been thinking about America's political division almost entirely as a political problem. Maybe I needed to understand it as an economic problem, too.

Because if anger keeps us watching, clicking, sharing and coming back... what is anger worth?

And if division has economic value, an even more uncomfortable question follows: Who gets paid when we hate each other?

​

I had assumed social media created most of this. It didn't. The Federal Communications Commission stopped enforcing the Fairness Doctrine in 1987. Rush Limbaugh's nationally syndicated radio program began the following year.

It would be too simple to say one caused the other. Changes in radio economics and technology also helped make nationally syndicated talk programming attractive.

But something important happened. Radio stations discovered that highly opinionated political programming could attract enormous audiences. And Limbaugh was spectacularly successful at it.

He didn't merely discuss disagreements over tax rates or federal regulations. His program turned politics into entertainment, personality and conflict. Liberals weren't merely people who reached different conclusions. They could become recurring characters, targets and punch lines.

The formula proved enormously successful, and conservative talk radio became a major force in American political media.

This wasn't an algorithm deciding what made people angry. It was something older. An audience deciding what it wanted more of — and a business learning to give it to them.

​

Cable news added pictures, panels, personalities and a nearly endless supply of conflict. Fox News developed a conservative-oriented audience and programming identity. MSNBC, particularly in its opinion programming, developed a liberal-oriented one.

And here's where I had to check one of my own assumptions. I expected to find evidence that only one side was doing this. That's not what I found.

A 2026 study examined Fox News and MSNBC mentions of every member of the U.S. House from 2019 through 2022. Researchers found that both networks presented distorted pictures of the composition of the two parties, generally with greater distortion when portraying the opposing party. The biggest distortion involved attention: members of Congress with heavy social-media engagement received overwhelmingly more coverage.

The study's author described it as the political “show horses” outshining the “workhorses.” Think about that.

The member of Congress quietly negotiating legislation may be doing something consequential. But somebody saying something outrageous about the other party? That's television.

There is an important warning here, though. Finding similar incentives on the political left and right does not establish that the content, scale, accuracy or consequences are identical.

“Both sides do it” isn't an explanation. Neither is “only the other side does it.” The interesting question is whether an attention economy rewards certain behavior regardless of ideology.

And that's where the story gets much bigger than cable television.

​

Radio hosts could hear the phones light up. Television executives could watch ratings. But social-media companies could measure almost everything.

And this is where my little observation walking past Fox News turned into something much bigger. Because researchers have actually measured what negativity does to our behavior online.

One enormous study examined more than 105,000 variations of online news headlines, generating roughly 5.7 million clicks from more than 370 million impressions. The researchers weren't merely asking people what they thought they would click. They were studying randomized tests involving different headlines for the same stories.

The result was remarkably straightforward. For a headline of average length:

​

Each additional negative word increased the click-through rate by 2.3%.

​

Positive language moved in the opposite direction: each additional positive word reduced the click-through rate by about 1 percent.

Think about what that means. You don't need a conspiracy.

You don't need an executive sitting in a dark room saying: Let's make America hate itself. You just need a dashboard.

Headline A gets fewer clicks. Headline B gets more. Do more of B. And eventually the system learns something about us.

We click on things that upset us.

This may be the part that disturbed me most. The machines aren't the only things learning. We are.

Researchers studying Twitter examined 12.7 million tweets from 7,331 users, then supplemented those observational studies with controlled behavioral experiments. They found that positive social feedback following expressions of moral outrage was associated with a greater likelihood of expressing outrage again. Their experiments also supported a causal role for social feedback, while the observational data showed how the pattern operates in real social networks.

Likes. Shares. Retweets. Attention. Approval. We learn what works.

The researchers also found evidence that people adapt to the expressive norms around them. In networks where outrage is common, outrage begins to look normal.

So the cycle doesn't require someone manipulating us from above. It can become self-reinforcing.

We express outrage. People reward us. We express more outrage. Other people see what gets rewarded. They adapt.

And eventually everybody begins shouting because shouting is what gets heard.

​

This part is almost difficult to believe. Beginning in 2017, Facebook's ranking system treated emoji reactions — Love, Haha, Wow, Sad and Angry — as stronger signals than a simple Like.

According to internal company documents later reported by The Washington Post, an emoji reaction was initially given five times the ranking weight of a Like.

The reasoning was that reactions represented deeper engagement. And engagement mattered.

But Facebook's internal research later found that posts producing angry reactions were disproportionately associated with misinformation, toxicity and low-quality news. The company subsequently reduced the weighting and eventually stopped treating Angry as a positive ranking signal.

But think about the lesson buried inside that episode. Nobody needed to program: SHOW PEOPLE THINGS THAT WILL MAKE THEM ANGRY.

The system could simply be told, in effect: SHOW PEOPLE THINGS THAT WILL MAKE THEM ENGAGE. And anger could help accomplish that.

That's a very different accusation. And, to me, a much more troubling one. Because you can fire a propagandist.

How do you fire an incentive?

​

This is where I started thinking differently about Fox News. And MSNBC. And political podcasts. And YouTube personalities. And TikTok creators. And, frankly, people like me who post political things on Facebook.

We tend to think of this as an ideological battle. Conservatives want conservative ideas to win. Liberals want liberal ideas to win. Of course they do.

But underneath that battle is another system that doesn't necessarily care who wins. It cares whether we stay.

Watch another segment. Click another headline. Read another post. Write another furious comment. Share another outrageous clip. Come back tomorrow.

Cable networks compete for viewers and subscribers. Websites compete for clicks and subscriptions. Influencers compete for attention. Social platforms compete for engagement and advertising revenue. Political organizations compete for donations and committed supporters.

Different businesses. Different politics. Different revenue streams. But they can encounter the same incentive: Don't let the audience become bored.

And few things are less boring than an enemy.

​

If the evidence stopped there, I could write a wonderfully tidy conclusion: Social media puts us in echo chambers. Echo chambers make us hate each other. Algorithms caused polarization.

There's just one problem. The evidence doesn't let me say that.

In a major 2023 study conducted with Meta, researchers examined Facebook use during the 2020 presidential election. For 23,377 users, they reduced exposure to content from politically like-minded sources by roughly one-third.

People subsequently saw more cross-cutting material and less uncivil language. But researchers found no measurable change across eight preregistered measures that included affective polarization, ideological extremity, candidate evaluations and belief in false claims.

That matters. Because it means I can't simply blame an algorithm for creating the country I see around me.

Maybe the relationship runs in both directions. Maybe the machines learn from us because we first teach them what captures us. And then they give us more of it. And then we respond to what they give us.

Human behavior shapes the machine. The machine shapes the environment in which humans behave.

Round and round it goes.

​

Remember what Facebook originally felt like? Your cousin's vacation. Somebody's new puppy. A picture of somebody's breakfast that nobody particularly needed to see. A kid's first day of school. A cat doing something ridiculous.

Plenty of that still exists. But there's an obvious problem if you're competing for human attention.

A kitten is cute. You smile. Maybe you hit Like. Then you keep scrolling.

But tell you that somebody is DESTROYING YOUR COUNTRY? Now you stop. You read. You comment.

Somebody disagrees with you. You answer them. They answer you. Someone else jumps in.

You come back twenty minutes later to see what everybody said.

From the standpoint of a system measuring engagement...

​

THE ARGUMENT JUST BEAT THE KITTEN.

I thought I knew the answer when I started. I'm less certain now.

The evidence doesn't support the simple story that Fox News, MSNBC, Facebook or any single villain created America's political divisions.

We bring something to this bargain, too. Human beings were tribal long before cable television.

We notice threats. We respond strongly to negative information. We like having our beliefs affirmed. And apparently, we sometimes enjoy being angry.

Media companies didn't invent those instincts. But modern technology became extraordinarily good at measuring them. And once attention became measurable, attention became something that could be optimized.

That's the distinction I hadn't really appreciated.

Maybe nobody had to sit down and decide to build a rage machine. Maybe thousands of people — hosts, producers, programmers, politicians, influencers and ordinary users — simply discovered that certain kinds of content performed better.

The incentives could do the rest.

​

I started with a television playing Fox News and the feeling that I was listening to a rage machine. I still think there was something real in that reaction. But I was asking too small a question.

The interesting question isn't: Why does Fox News make people angry? It's: Why is anger so valuable?

And the answer seems to involve some uncomfortable combination of human psychology, political identity, technology and economics.

Negative language can attract more clicks. Social approval can reinforce expressions of outrage. Partisan television can disproportionately show us the loudest and most socially prominent members of the political tribes. Algorithms can learn that emotionally provocative material generates engagement even when nobody explicitly tells them to seek anger. Businesses can benefit from the attention that follows.

But the research also warns against the easiest conclusion: simply reducing people's exposure to politically like-minded content did not, in one large Facebook experiment during the 2020 election, measurably reduce their political polarization.

None of this proves that every angry television host, social-media executive, politician or influencer secretly wants Americans divided. They don't need to. The incentive can exist without the conspiracy.

And maybe that's the part I didn't understand before.

​

If reconciliation produces less engagement than outrage, who in the attention economy is financially rewarded for reconciliation?

If social-media algorithms stopped rewarding emotionally provocative content tomorrow, how much of our political hostility would remain?

Have Americans actually become as hostile toward one another as our screens make us appear — or are our screens disproportionately showing us the angriest people among us?

How much responsibility belongs to Fox News, MSNBC, Facebook, YouTube, X and TikTok — and how much belongs to us, because we're the ones who keep clicking?

Can a news organization remain profitable while regularly telling its audience things it doesn't want to hear?

Can an influencer build an enormous following by repeatedly saying:

The other side has a reasonable point?

And perhaps the question bothering me most now isn't the one I started with.

I started by wondering why that television sounded so angry.

Now I'm wondering what would happen if it didn't.

Because if outrage attracts attention, if attention can be converted into value, if enemies are more engaging than neighbors, and if media companies, algorithms, politicians, influencers and audiences can all participate in the cycle...

Who makes money when America calms down?

“I don’t know.”