The Business of the Culture War

How Activists Undermine Their Goals

The world seems to be angrier these days. Everyone’s concerned about cultural issues. If you turn on the news, you can flip over to Fox News and hear about the wokes, the transgenders, and the left-wing liberal media; or go over to MSNBC and hear about book bans in the school library. It is rare to see discussion of substantive economic issues on television. Instead you hear about disasters, crime, and one outrageous incident a day.

But the politicians aren’t that angry in their messaging. The culture war is not what you run a campaign on, unless you are in a safe district and can indulge in it. Instead, politicians run on pocketbook issues. Their ads are warm and pleasant, and tell you about how they will work to cut taxes, lower costs, and put money in your paycheck.

I presume that they are not making a mistake. This must be a winning strategy for each. But why is it the case that politicians don’t win by ginning up outrage? And why does this explain why the Democratic Party is so afflicted by “The Groups”?

I had the pleasure of reading “The Business of the Culture War” by Shakked Noy and Aakaash Rao recently, which pretty much explains everything about the modern world. Their core insight is that media companies and politicians have different motivations. Media companies want to sell advertising space, and so care only about the raw number of people watching. Politicians do not care about the number per se, but only their number relative to the opposing candidate. A politician benefits doubly when they persuade a voter to change their vote, while a media company benefits from a message that makes people so angry they have to watch, even if it leads more people to view their competitors.

To make this case, Noy and Rao need to collect data, and establish a few facts. First, they have details on advertisements created by all candidates for major offices since 2000, plus presidential ads from 1960, and the complete video recordings of every major news network since 2014 plus transcripts going back to 1968. To deal with this extraordinary quantity of data, they have an LLM characterize segments as dealing with very narrow topics, which can then be categorized themselves. The big stylized fact is that culture war topics are way more common in the media than in ads.

The ads of politicians of all kinds have much more economic content than television does, though this has trended down, and in recent years broadcast television has converged with cable news.

But why do cable news companies broadcast this content? Now we need details on viewership second-by-second, which, of course, they have. They can now test which types of content gets viewers by exploiting plausibly exogenous variation in when someone flips onto a channel. While people can anticipate what programming is on at any given minute to some extent – you have the headlines at the top of the hour, human interest at the end, and so on – you can use unexpected variation in what content is on relative to what is normally on as an instrumental variable. If they flip on, and stay, that content is effective in getting voters. You can also track what content is on when viewers change the channel.

The results are as we expect. Nothing is as strong as flipping to a channel and seeing advertisements, which absolutely tanks viewership, but economic content performs worse than cultural content. Importantly, the viewers gained to not come from other news channels, but from either watching non-news programming or just turning off the TV. In fact, when people change the channel when economic issues come up, they are less likely to go to the other news companies than to other entertainment options.

It is important to note that these are just the short-run effects on viewership. Cable news companies need not profit by switching over to only culture war topics, because they could value the reputation of actually telling some of the news, some of the time. The gap between broadcast and cable news is consistent with broadcast news being something aired to make the network respectable, and the recent convergence due to firms finding out that trashiness sells.

We also need to argue that the politicians are correct in thinking that pocketbook, kitchen-table issues make them more likely to win. This part is not going to be as credibly identified as the cable news portion, because politicians aren’t strictly maximizing their share votes, but rather their positions weighted by their chance of winning, and politicians are in turn chosen by voters who are trying to solve the same problem. You should be concerned that choice of ad is correlated with unobservable features like “charisma” – you can imagine a candidate who knows they are going to win, so they put out ads which appeal to activist groups so that they are better positioned for leadership positions in the House when they inevitably get there. In particular, raising donations in a safe seat allows you to transfer them to swing state candidates in exchange for good will, and this is something where you do not particularly care about the voters in your district who might be mobilized to vote against you.

Noy and Rao’s approach is to use two-way fixed effects, controlling both for the candidate and the year. (Some candidates are more charismatic than others, so we assume that this is constant across years and adjust. Likewise, some years are better for one party than another, so we assume it is constant across candidates and adjust). Candidates which air more economic ads tend to do better. There is selection on observables, which we controlled for – you can say that selection on unobservables would have to be substantially larger than observed selection in order to explain away the results entirely. Another piece of suggestive evidence is that the swing in voting due to ads comes from higher propensity voters, which suggests that they are persuading people who are politically active.

Finally, we need some evidence that all this cable news stuff matters. We’re likely to underestimate this, because “culture” will spill across treated and untreated units, but they can reuse instrumental variables for plausibly exogenous exposure to cable news to see how it changes the priorities of potential voters. They use the Martin and Yurukoglu (2017) instrument of channels being assigned higher or lower channel numbers in a way that seems random. Voters who live in places where they are exposed to more culture war programming due to the channel number placement are more likely to say that their most important issue is a culture war one. Politicians pick up on this, and include more advertisements focusing on cultural issues.

This tracks with prior work on the effect of Fox News. Stefano Dellavigna and Ethan Kaplan (2007) use the rollout of Fox News to different parts of the country to argue that it caused 3 to 28% of its viewers to vote Republican who would not otherwise have voted or voted for that party. Their argument for identification is that each small town has only one cable provider, but who exactly this cable provider is varies in ways that are not correlated with observables and appear to be somewhat arbitrary. Martin and Yurukoglu, mentioned above, found that Fox News changed vote share, while MSNBC did not.

But this paper says so much more about the world than just the media. The incentives of private organizations which are involved with politics are totally misaligned from those who are actually trying to win elections. If you are an organization which cares only about the influence of your particular organization, not the issue you are focused on as a whole, then you will behave in ways that grab attention even as they make your own views less likely to be the law.

We can thus see at once why protestors so often behave in ways that are alienating and ineffective. I cannot possibly believe that throwing soup cans on paintings has done much at all for climate change. If a vote was called on whether people approved of Just Stop Oil, they would likely lose it by much larger margins than when they started the campaign. Yet, an organization’s real fight is not for approval, but against potential supporters not knowing who they are. In that, they have been extremely successful.

This is just like how firms compete both by finding technologies, and by stealing business from their rivals, as we covered earlier this week in talking about Aghion and Howitt. Having lots of groups leads them to take actions which parasitize each other, at the expense of the party as a whole. The only way out is to impose discipline – we need to punish groups which behave in ways that sabotage the party for their personal benefit.

Aakaash Rao is on the job market this year. All I can say about him is that when I saw this paper be presented, a ton of grad students in the lower years turned up to support him. He had been their TA, and was universally well-liked. Consider that an endorsement of things besides his paper, which needs no elaboration.