What Are We Betting on in Politics?

By Brandon Tatum, CEO of the National Governors Association, Taubman Center Visiting Fellow College football season is about to get underway, which means I already know what is coming in my house. My 12-year-old will start trying to make friendly bets with me about games. Who wins? By how much? Who scores first? There is something fun about making a prediction and then seeing whether you were right. Maybe that is part of the reason I have become so interested in prediction markets. Over the past several months, I have developed a weekly ritual of looking through the political markets to see what they are telling us. Most of the attention around prediction markets focuses on the obvious questions: Who is going to win an election? Which party will control Congress? Will interest rates go up or down? But I have become increasingly interested in whether these markets can tell us something less obvious about the political environment. Recently, I noticed a category of markets that asks people to predict the actual words and phrases political leaders will use in upcoming speeches, interviews and public appearances. That caught my attention because I have also been spending a great deal of time researching dignity and contempt in American politics. It made me wonder whether prediction markets could help answer a different kind of question. We have plenty of polling that tells us what Americans say they want from their political leaders. Pew Research Center found that 78 percent of Americans believe elected officials should avoid heated or aggressive language, including 83 percent of Democrats and Democratic-leaning independents and 75 percent of Republicans and Republican-leaning independents.1 And some remarkably fresh research released in June 2026 gives us an even more detailed picture of an electorate that is considerably more complicated than the red-versus-blue caricature that often defines our politics. Pew's new political typology, based on more than 10,000 Americans, identified nine distinct groups based on political and cultural values.2 This research is barely two months old. It offers one of the freshest looks we have at the American electorate, and it suggests that the public is considerably more complicated than our political discourse often allows. One of Pew's groups is even called the “Pragmatic and Polite Right,” a group characterized in part by its preference for civility, cooperation and bipartisan common ground. All of this tells us something about what Americans value. But prediction markets give me an opportunity to look at the question from the other direction. Instead of asking what people want politicians to say, what happens when people have to predict what politicians are actually going to say? That distinction matters because what we value and what we expect are not necessarily the same thing. A traditional poll can ask whether a political leader should treat an opponent with respect. A prediction market asks participants to make a judgment about what they believe that political leader will actually do. One gives us insight into our values. The other may give us a signal about our expectations. So I decided to take a closer look at what the markets were actually predicting. Were people expecting the kind of political behavior they say they want, or had we learned to expect something very different? I reviewed 36 forward-looking political speech contracts and classified the language into three categories: contempt and conflict, neutral or policy language, and dignity or cooperation. The sample is small. It is not a nationally representative survey, and prediction-market traders should never be confused with a representative sample of the American public. I would not draw sweeping conclusions from 36 contracts. But the initial results were interesting enough that I think the question deserves considerably more research. When contempt or conflict-oriented language appeared in the markets I examined, traders assigned it an average probability of approximately 52 percent. Language associated with dignity or cooperation averaged approximately 46 percent. Neutral and policy language averaged approximately 38 percent. The six-point difference between contempt and dignity is not the story. The sample is far too small for that. What interests me is what these markets allow us to observe: expectations. When traders were asked to predict whether political leaders would use language associated with conflict or contempt, they viewed that behavior as entirely plausible. That is very different from endorsing the behavior. Someone who expects a politician to insult an opponent is not necessarily saying they want the politician to do it. They may dislike the behavior entirely. They simply believe it is likely to happen. Americans can value dignity while expecting contempt. They can want cooperation while anticipating conflict. They can believe political leaders should behave differently while simultaneously believing those leaders probably will not. There is nothing contradictory about those positions. In fact, the distance between what we value and what we have learned to expect may tell us something important about the condition of our politics. But that distance also raises another question. If our expectations can move in one direction, can they move back? If political conflict has become predictable, what would it take for dignity to become predictable too? That is the question I find more interesting than the initial prediction-market numbers: What would have to happen for the market to flip? What would need to change in American politics before people began expecting dignity rather than being unsurprised by contempt? There is reason to believe such a change is possible. Research associated with Stanford's Polarization and Social Change Lab challenges the assumption that political incivility is always an effective strategy. Researchers found that politicians who attack their opponents can damage their own reputations in the process. Their analysis of congressional language similarly found greater civility associated with higher approval, while experimental work found that civil political communication produced higher approval among most of the groups studied.3 This creates an interesting disconnect. Polling suggests Americans value less aggressive political behavior. Research suggests civility can actually be politically effective. Yet our expectations have not necessarily caught up with either finding. We have seen conflict often enough that it has become easy to predict. Perhaps that is the more useful way to think about the prediction-market numbers. The important finding is not whether contempt currently has a six-point advantage over dignity in a small sample of contracts. It is whether that expectation can change. Prediction markets move when people receive new information and begin believing a different outcome is more likely. Political expectations should be capable of moving in much the same way. If political leaders consistently demonstrate that they can win elections without humiliating opponents, solve problems across party lines and remain strong without relying on contempt, people may eventually adjust their expectations. Dignity would no longer be something Americans simply say they value. It would become behavior they reasonably expect to see. That gives us a different benchmark for improving political culture. We should certainly care whether Americans value dignity. But perhaps we should also care whether they believe their leaders are actually capable of practicing it. Maybe the goal is to get us to a place where dignity is worth the bet. If the market opened tomorrow, would you put $100 on dignity? 1 Pew Research Center, “Nearly 8 in 10 Americans Say Politicians Should Avoid Heated or Aggressive Speech,” November 25, 2025. The survey of 3,554 U.S. adults was conducted August 4–10, 2025. Seventy-eight percent said elected officials should avoid heated or aggressive language, including 83 percent of Democrats and Democratic-leaning independents and 75 percent of Republicans and Republican-leaning independents. 2 Pew Research Center, Beyond Red vs. Blue: The Political Typology, released June 10, 2026. Pew surveyed 10,357 U.S. adults and classified the public into nine groups based on 30 questions about political and cultural values. The report also draws on additional surveys of the same respondents through April 2026. 3 Jeremy Frimer and colleagues' research on political civility, discussed by Stanford's Polarization and Social Change Lab. Their work finds that political attacks can impose reputational costs on the attacker, that greater civility in congressional speech is associated with higher approval, and that experimental exposure to civil rather than uncivil political communication generally produces more favorable evaluations.