Nationalized Elections, The End Of Local News, And Government Accountability

When was the last time you voted split-ticket in an election? It may not be surprising to hear that our elections have become increasingly nationalized in the last few decades. Most people vote for a single party straight down the ballot. The question is, why?

Daniel Moskowitz, Assistant Professor of Political Science at the Harris School of Public Policy, says the answer may be the massive reduction of local news in his recent paper. On this episode, we speak with Moskowitz about why nationalized elections are a problem, the key role of local news, and what we might do to fix things.

Listen on Apple Podcasts or wherever you enjoy podcasts.

Transcript 

Wioletta Dziuda:

I'm Wioletta Dziuda.

Anthony Fowler:

I'm Anthony Fowler.

William Howell:

I am Will Howell, and this is Not Another Politics Podcast.

William Howell:

When we think about what's happened over the last 20, 30 years in the media environment, there've been these extraordinary changes. We've seen this explosion of social media, the rise of new news sites online, increasing subscriber base for some national print outlets, Fox News. But amidst all this is a kind of decimation of local news. It's really one of the most striking developments of the last, I think, quarter century, when we think about the ways that average citizens learn about politics.

Wioletta Dziuda:

Yeah, and I have a feeling that it's harder and harder to learn about what's happening in your city or in your state or in your county. Recently in Chicago here, we had this huge controversy where the mayor Lori Lightfoot was negotiating with the teachers union whether to open the schools or not, and I really tried to figure out what's going on. I tried to form an informed opinion whether I should support the Lori Lightfoot stance and perhaps vote for her in the next election cycle, or whether I should actually really stand by the teachers. I had a hard time finding article that would cover this issue in depth, that would explain what the positions of the different parties are. It was extremely frustrating.

William Howell:

Yeah. I mean, you spend your life thinking about politics and writing about it, and here you are trying to make sense of local events, local politics, and aren't able to collect the most basic information that you need because of the decline of local newspapers and the adequacy of their coverage. So, boys, it's like this is tied to profound concerns about voter accountability and the health of our democracy and the extent to which we can hold elected officials account for the positions that they take and their performance in office.

Anthony, you talked to one of our colleagues, who has a paper on precisely this topic.

Anthony Fowler:

I did. I talked to Daniel Moskowitz, who just published a paper in the American Political Science Review, and it's on precisely the effects of exposure to local news. To the extent that local news affects people's voting behavior, this might help us understand what's changed over the last few decades as local news has declined in its reach and its popularity.

Anthony Fowler:

All right. I am really happy to talk to my colleague, Daniel Moskowitz, assistant professor at the Harris School, who has written a lot of interesting papers about elections and representation. In particular, we're going to talk today about the nationalization of US elections. So what do you mean when you talk about the nationalization of US elections?

Daniel Moskowitz:

When people talk about the nationalization of elections, usually they mean that electoral outcomes are more closely tied to presidential election outcomes. That is, they're voting straight party ballots more often than they used to. So, what could drive that kind of behavior? Polarization, and that is that candidates within the same party are more similar to one another than they used to be. That if we go back to the '70s and '80s, a lot of Southern Democrats held very different policy positions than Northern Democrats did, so it wouldn't be unusual that Southern voters might split their ticket. That being said, we've observed this outside of the South still at a very fast rate, so we can't just attribute it to a Southern realignment that's going on. That's certainly one explanation, and it's more of a elite driven model of what's going on.

Another potential explanation, which is the one that I sort of get at in the paper is what people might call information or the nationalization of news. Voters have less information about candidates down ballot than they used to, because we've experienced a massive upheaval in the news environment over the past 20 or 30 years. There's been a huge decrease in circulation of local newspapers. Employment at newspapers has declined enormously, 40% or so. To some extent, local television newscast audiences have decreased, though there are more hours of local television news than there used to be. But a consequence of the upheaval in the news environment is that it's probably more difficult for voters to get information about candidates for down-ballot offices, and a result is that they might just apply their assessment at the top of the ticket to these down-ballot races.

Anthony Fowler:

So when it comes to measuring the nationalization phenomenon, what you focus on in your paper is split-ticket voting. How do you measure split-ticket voting, especially given that we have a secret ballot in the US and we don't know which candidates each person voted for?

Daniel Moskowitz:

There are a couple of ways of going about doing it. One is, which is primarily what I do, is to focus on survey data. So, individuals fill out these surveys about what happened in the past election, and they report who they voted for, for not every office on the ballot, but the major offices on the ballot, including House, Senate, governor, president. An alternative, which I also do as an extra check in the paper, is to rely on precinct-level data and use a precinct-level proxy measure split-ticket voting. There are some rare states, like South Carolina, for which they release ballot level data as part of the election audit. What I was able to do is I was able to look at what the true rate of split-ticket voting was within precincts, and then compare it with my proxy-level measure at the precinct level. I saw a very high correlation, so that kind of added some validity to the use of the precinct-level measure. But it makes interpretation of the quantities a little more difficult than using the survey measure of split-ticket voting.

Anthony Fowler:

Give us a sense of, as far as we know, how many people in America do cast split-ticket votes, and how has that changed compared to historic levels?

Daniel Moskowitz:

In the 2012 and 2016 elections for president/senator split-ticket voting and president/governor split-ticket voting, about 8 to 9% of individual who cast a ballot in 2012 and 2016 cast a split president/senator or a split president/governor ballot. Historically, that is much lower than it used to be. For president/senator split-ticket voting, in the '70s and '80s, it was around 25% of voters cast a split president/senator ballot, to give a sense of how much it's declined. It's about a third of the rate that it used to be when it was at its peak. I think it's a pretty similar decline if we look at president/House split-ticket voting in that timeframe. It was also about 20% in the '70s and '80s, and then kind of declined, starting in the '90s and 2000s, and is below 10% today.

Anthony Fowler:

You mention as one explanation for this, this informational story, that maybe there are a lot of people who would vote split ticket, maybe they would vote for Joe Biden for president and a Republican for governor, let's say, but they just don't even have very much information about the gubernatorial candidates in their state, so they end up... They vote for the Democrat because they usually agree with Democrats, even if maybe in some cases, if they looked into it, they would switch into the other way. So your paper's largely about that story, and you have a pretty, I think, innovative and clever way of getting at that story. The focus is on local news. Maybe some of our listeners, who... If you're a young person, you might think that local news is dead and doesn't matter. Is local news still relevant to the American politics today?

Daniel Moskowitz:

I think the answer is yes, it's still relevant. Local television news in particular is less subject to some of the economic pressures that local newspapers have faced, and that's just because people are still subscribing to cable. In your cable subscription fee, there are retransmission fees that go back to local stations in order for that cable provider to include them in the channel lineup. That helps fund their local news. As it turns out, the FCC also has rules that require local over-the-air stations to provide a certain number of hours of local television news. The media environment and the economic environment is very different for local television stations than it is for local newspapers.

But these local newspapers still do exists, even though lots of them have either gone away or are a shadow of their former selves. Strong, local newspaper in a media market, local television news benefits from that enormously because they do a lot of the deep investigative reporting that then local television newscasts will feature that reporting. So it still plays an important role, but I would say that it plays less of a role than it used to.

Anthony Fowler:

What sense do we have of just the ratings and the numbers? How many people still watch their local news coverage on television with some regularity?

Daniel Moskowitz:

Local television newscasts still actually have pretty amazingly high audiences, despite some decline in the audience of the late night newscast, which is traditionally the most watched news slot. I think that has declined about 30% over the past decade. But nevertheless, if you look at recent Nielsen numbers, local news reached about 40% of working-age adults in a typical week in the first quarter of 2017. These people watched on average about 2 and a half hours of local news in that week.

Anthony Fowler:

Tell us a little bit more about your research design. How do you go about estimating the effect of access to local news? That sounds like a hard problem.

Daniel Moskowitz:

Yeah. Media exposure is really difficult to study, because people select into different sources of media. If I told you, for instance, that people who choose to expose themselves to Fox News are more conservative, I think both of us would be very skeptical to attribute their conservatism to exposure to Fox News. Consumers choose things, and they choose things based on their preexisting preferences. So how do we know if the thing they choose to consume is actually causing any change in behavior or change in preferences, et cetera?

I was seeking out some sorts of variation in exposure to local television news that people didn't get to choose whether or not they watched it or had access to it. What I did was I kind of used an empirical strategy that lots of others have used. I used the geography of media markets. Today's media markets are mostly based on where television signals travel, way back when television first rolled out. That resulted in the definition of media market boundaries. Due to various rules from the FCC as well as exclusivity contracts between networks and stations, those media market boundaries haven't changed much over time.

What it means is the media market you live in determines the access to what stations you have. Some of these media market boundaries, because they're idiosyncratic, cross state boundaries. So the Fort Wayne, Indiana, media market goes into Ohio. Two Ohio counties are contained within the Fort Wayne media market. But 93% of Fort Wayne's media market population resides in Indiana and only 7% resides in Ohio. So if you happen to be a resident of those two Ohio counties, it means that you're getting basically Indiana television stations. The coverage of those Indiana television stations tends to focus on Indiana office holders: the Indiana governor, the Indiana senators, et cetera. So it's much more relevant to the audience that lives within the state of Indiana than those 7% of that media market population that lives in Ohio.

So what I did was I used closed captioning data. The closed captioning data allowed me to look for mentions of office holders from different states from these stations in their local newscasts. So I looked for mentions of the governors and the senators in each of the states that have any population overlap with the media market. What I found was that if you look at someone who lives in almost entirely out-of-state media market, that is a media market for which almost all of the residents, all of that media market are located in another state, they get about a half a mention per hour less of each senator and about 1.5 fewer mentions of their governor per hour of non-entertainment television.

Anthony Fowler:

And then you also check within the survey data to see if it looks like, indeed, people in these in-state media markets actually do know more about their senators and governors.

Daniel Moskowitz:

Yes. I used some what I call knowledge measures about those senators and governors, based on the same survey data that I used to assess whether they engage in split-ticket voting. I look at whether they are able recall the party of the senator and the governor, whether they are able to place them on an ideological scale, say they're very liberal or somewhat liberal, moderate, very conservative, somewhat conservative, and whether they're able to evaluate their senators and governors, which is just to say whether or not they approve of them, with the idea being that if people say they don't know how to place them or they're not able to evaluate them, that means they don't have enough information about those office holders to be able to do so.

I find very large effects of residing in an in-state media market on these knowledge measures. For instance, for the ability to recall the party of the senator or governor, people who live in an in-state media market are about nine percentage points more likely to recall their senator and about 10 percentage points more likely to recall the party of their governor than someone who lives in an almost entirely out-of-state media market.

One thing I also tested, to kind of confirm the validity of the research design, was whether there were differences in general or national political knowledge. Because if it just happens to be the case that the people who reside in these counties are different from the people who reside in the in-state market counties, then we might expect them to just be more knowledgeable about political in general, not about their office holders in particular. So I looked at a bunch of knowledge of party control of the House and the Senate, ability to place the parties on an ideological scale, placing the Democratic Party to the left of the Republican Party. I didn't observe differences for people who reside in out-of-state markets relative to people who reside in in-state markets on these general and national political knowledge measures, which was really convincing, to me at least, that the research design has some validity and that the results are not likely driven just by unobserved differences between residents of in-state markets and out-of-state markets.

Then, when I took the next step to look at, okay, so I observed differences in terms of political knowledge, does that then have an actual effect on voting behavior in the voting booth? So I looked at casting a split ticket ballot for a president/senator and president/governor. What I observed is that compared to people who live in almost entirely out-of-state markets, that is in a county for which their state makes up a very small share of the media market's population, people who reside in in-state markets are about two to three percentage points more likely to cast a split president/senator ballot and people who live in an in-state media market are about four to five percentage points more likely to cast a split president/governor ballot. These are large effects when you consider that only 8 to 9% of people are casting a split president/senator or split president/governor ballot.

Anthony Fowler:

Are there any obvious solutions to that problem? To the extent that local news is declining and to the extent that a lot of voters aren't getting information about their governors and senators and so forth, and to the extent that that's leading to the nationalization of elections, are there ways of addressing that problem or mitigating that problem, if we think that this is hurting electoral accountability and hurting representation and so on?

Daniel Moskowitz:

I'm always hesitant to draw normative conclusions, because if we go back not that long ago, there is an APSA kind of taskforce where panels were put together that basically bemoaned that the parties were indistinct from one another, that incumbency advantage was too large, that there are all these ills of American politics, that now we look back at those times and we're like, "Gosh, that looked a lot better than it does now." When we now have pretty distinct parties from one another and it's resulting in all these gridlock and lots of other issues that we're very concerned about. So it's much easier to diagnose your present ills than it is to forecast what would happen if things changed, because things might not be better if we do something. There are lots of unforeseen consequences.

In terms of what to do about the media environment, there are, at least from my vantage point, lots of good solutions when it comes to newspapers. There's lots of complaints about the acquisition of local newspapers by these large chains or private equity firms that are cutting the reporting budget, staff, et cetera, which is making it more difficult for these newspapers to cover their local areas, their officials. What is the solution to that, such that these newspapers have the resources they need to function properly? Well, it is not clear what the alternative is other than some sort of non-profit model in which you just have these wealthy benefactors who are funding it out of the goodness of their heart, which doesn't make me feel better about the situation, because the benefactors could just as easily be evil rather than benevolent and just try to get information out there to people in a way that is not objective or unbiased or whatever other standards we hope our newspapers strive for. So I don't have any good sense of what economic models exist that could allow local sources of news to function in a way that we think is healthy for our democracy and is also sustainable in terms of their finances and keeping them running.

In terms of local television news, there are maybe some small steps that the FCC, the Federal Communications Commission, can take to preserve what they call localism in broadcast news. The previous FCC commissioner took some steps that were somewhat controversial, at least from the perspective of people who really want to preserve local broadcast television news. So one of the things they did was they eliminated what's called the home-studio rule, which used to require that stations have a studio located in their local media market. The thought behind it is if you get rid of that, then it makes it easier for a conglomerate of stations or any sort of company that owns lots of stations across various media markets to pipe in more nationalized programming into those local media markets. If you have reporters on the ground, in the studio, then that might mean that you get more and better local coverage. That was one change they made that Biden's commission FCC could potentially revert back to.

They also lifted a cap that restricted the number of stations that a single entity could own. In other words, they used to only have so much of an audience reach. Because of this cap, you could only reach some percentage of the American population that a single entity, the stations that they owned, could reach. That was lifted, and that allowed the rise of potentially your Sinclairs, et cetera. It actually could have been further eroded, but to the surprise of many experts, Trump's FCC did not allow a further expansion of Sinclair. But there is a lot of concern about this idea that a single entity can own stations across the entire country or large swathes of the country, and then they can pipe in more nationalized programming and less local programming for those local television newscasts. So there are certain restrictions that the FCC could take to require that the content of these newscasts needs to be sufficiently local.

William Howell:

Can you give us a two-minute encapsulation of what the thrust of the paper is and the key takeaway points are?

Anthony Fowler:

It's a simple design, but it's a very clever one and a very useful one, which is he's looking at how voters vote in the same election, in the same state, but some of those voters happen to live in a media market where most or not all of their media market is in that state, so all of their local news coverage is going to be about that state. If you live in Indianapolis, local news coverage is going to be what's going on with the Indiana governor and the Indiana senators and so forth. If you live in north-west Indiana, you're going to be getting Chicago television. He finds that the people in Indianapolis are much less likely to vote for the same party across offices than the people in north-west Indiana, presumably because, even though they might typically be a Democrat, let's say you supported Joe Biden, you might find out that actually this year the gubernatorial candidate from the Republican side is really experienced or really competent or appealing in some other way, and you wouldn't have known that if you lived in north-west Indiana.

Wioletta Dziuda:

For me, the paper is a good-news-bad-news story. The good news is that it's not what many people actually say in the media or even in political science, that people just vote based on the party affiliation. They think of themselves as Democrats the same way they think of themselves as supporting Red Socks or whatever your favorite... I don't know nothing about American sports.

Anthony Fowler:

The sports team, not the clothing item red socks. You have to...

Wioletta Dziuda:

FC Barcelona. But there is this narrative in the media and also in political science that voters just treat parties as sports teams, and no matter what you do as a politician, you are not going to be rewarded or punished for that. It all depends on whom you're affiliated with. In a sense, Daniel's paper is good news. The bad news is, unfortunately it seems that people are getting less and less of this information, so they behave as if politicians and parties were just sports teams.

William Howell:

Which is... Anthony's favorite notion of the voter is that they're just casting votes blindly, out of naïve allegiance to their favorite sports team. But why should we care about split-ticket voting?

Anthony Fowler:

Sure. No, that's a good point. I'm sure many of us have voted for the same party in lots of different elections, and we don't think we did so because we were unsophisticated or stupid or naively brand-oriented or something like that. But the troubling pattern is, if it looks like voters always vote for the same party, no matter what, then what incentive to politicians have to actually do a good job? So you can understand why people are concerned about this increasing nationalization of election and nationalization of voting behavior and decline in split-ticket voting. One reason we care is that we want people to actually try to figure out which candidate is better, and every once in a while that might be the other party that you typically don't support.

William Howell:

One way to understand what these effects are is that with this additional information, voters are evaluating candidates differently, and on the basis of those evaluations, they're casting votes that differ. Alternatively, what you might say is that elected officials, recognizing the systematic differences in the extent to which their electorate is receiving information about them, are altering their behavior. So if I happen to have my base in north-west Indiana, and I'm a governor of Indiana, and I say, "Look, my base isn't getting as much information about me," that might induce me to take a set of positions or behave in a way that differs, and it's that change in behavior, which again is not observed, is what induces then the split-ticket voting.

Wioletta Dziuda:

No. No. No. Because I don't... No.

William Howell:

Not at all.

Wioletta Dziuda:

No. I get the first part of your story, that if you know that your voters get little information about you, you would presumably behave differently. You have less incentives, perhaps, to exert effort and more incentives to engage in corruption because the outcome of the election is not going to depend on those actions. But how does this feed back into voting straight ticket, I don't see.

Anthony Fowler:

I buy what Will is saying, I think. I can try to connect the dots, maybe. Daniel's paper is more about the voters, but I think the implications for politicians are really important, because what we care about is are good policies being made and so forth. So think about this comparison: if you are the governor of Utah, the Salt Lake City media market more or less is the same as the state of Utah. Not exactly, but pretty close. So, everybody who lives in Utah for local TV coverage, they're learning what you're doing.

If you're the governor of New Jersey, nobody in New Jersey is getting a lot of local television coverage about New Jersey because you either live in the Philadelphia media market or you live in the New York City media market. So if you're the governor of New Jersey, you have much less incentive to work hard, do a good job, implement policies that the voters want, because nobody will even find out if you do so you might as well slack off a little bit, engage in a little bit of corruption and so forth.

It probably is fair to say there probably is a lot more corruption happening in the New Jersey state house than in the Utah state house on average, and maybe that's even one of the reasons, although that's not a great study that I just did. But you see the idea. So the implication of Daniel's story is very important for the actual policies that get implemented where you are. And that also relates back to how people vote, because of course, when people do a bad job in Utah, the voters will vote against them, and when the governor does a bad job in New Jersey, the voters often won't know and they'll just vote for the same party they vote for usually in presidential election.

Wioletta Dziuda:

Yeah. So you are saying that in situations where there's very little accountability, the reason for straight-ticket voting would be the voters just have no reason to vote in favor or against a particular candidate because they have very little information, so they just fall back on party cues. There is that slight emotional affiliation with a particular party. That could be. That could be.

Anthony Fowler:

And of course, there are studies that focus on precisely this kind of thing. They use the same idea. There are studies that look at differences across congressional districts, in terms of how well aligned are congressional districts and newspaper media markets, for example. You do find that when the congressional district doesn't align very closely with the newspaper media market, then the member of Congress tends to vote in a more extreme manner or not show up to committee hearings more often and so forth. When there is a close alignment, they work harder and they do a better job. So these kinds of things do matter for the way that voters are represented and the policies that get made.

Wioletta Dziuda:

One question we could try to think about a little bit is the numbers that Daniel is getting, and are these numbers high enough to make us worry about accountability. Because at the end of the day we don't care that every voter in New Jersey knows about what the governor is doing. It's fine if some fraction of the population votes always Republican and some fraction votes always democrat, as long as we have enough voters in the middle who actually pay attention to the news or are informed. Perhaps we can't really look at his numbers too carefully, but what he told us is that around 8% people do seem to be responsive to some information, because they do engage in split-ticket voting. The access to good media market seems to increase that number by four percentage points. So it seems like a huge increase, but perhaps already this 8% who are switching their vote based on information is still enough to provide accountability. I think this is a question that definitely is not answered in this research, and it's a very interesting question.

Anthony Fowler:

There's a few things to say about that. I mean, one, Daniel's estimates don't really tell us what share of voters really care about the performance of the governor and vote accordingly. It could be the case that most of the time, just by using your normal partisan leaning, you'll get the right answer, in the sense that even if you're more informed, that's probably the candidate you would have voted for anyway.

I mean, it must be true more often than not, that maybe if I don't pay very close attention what's going on in the Illinois gubernatorial race and I just vote for my normal party, I'll be right more often than not. So in some sense, the fact that I get a, say, two or three, four, percentage point effect depending on the specification, that's a lower bound on the number of people that really care. Of course, there are a lot more people who probably would have voted differently had the information been different. Had their partisan aligned governor actually been doing a really bad job, many more people would have potentially responded to that information. So that's one reason why the numbers are a little bit hard to interpret. It could be the case that, in fact, there are lots and lots of people who respond to this information, and that's enough of an incentive for the governor to actually do a good job and not engage in corruption and not slack off and so forth.

Then there's the other complicated part of this, which is that just because there's a vote-share effect, doesn't mean that that's actually tipping a lot of election results. Illinois is a pretty Democratic state. Even if you way underperform, the Democratic Party, it's typical performance in Illinois, the Democrat is still usually going to win. So maybe there's still not a whole lot of incentive for somebody like J.B. Pritzker to work really hard and avoid corruption, because he'd have to be astoundingly bad to lose reelection. I mean, they do lose. Democrats have lost in Illinois. But that's another complex part of this story that you have to think about.

William Howell:

You point to how close you are to the margin, right? Are the effects big enough to somehow tip an election? There also are concerns about the extent to which changes in split-ticket voting on one side may well be offset by the other side, that is the incidence of split-ticket voting could go to the roof, but there are offsetting effects. The votes that standard Republicans give up to Democrats are offset by a rise in votes that usual Democrats now are given to Republicans, and you'll get the same electoral outcome.

Anthony Fowler:

I was just going to say that we do know that an average, there are lots of these effects. On average, more ideologically moderate candidates do better. We know that voters respond to the economy, for example. So, we know that on average there are things that lead voters to change their votes in a systematic direction. It's not like the least informed the voters are, the more they're able to take whatever their preferences are and translate them into votes, and there is going to be some on average systematic effect here.

William Howell:

I don't know. I think it varies. Like I said, it varies. Let me give you this example, it doesn't involve a candidate, but it involves learning about charter schools. Here is a specific issue. You could learn more about it, like what is a charter school? Most people don't have a clue what a charter school is. It turns out, when you inform citizens... Let me tell you that they are in fact public schools. They're funded by the government. And prayer is actually not allowed in them. That has the effect of increasing support for charter schools among liberals and decreasing support among Republicans, but on average, the level of support that you observe doesn't shift very much one way or another, even though there are big changes within each groups. This is like a case where more information, you start seeing changes within parties in their willingness to support this given policy, but that they're offsetting. When you think about tabulating votes, if they are offsetting...

Anthony Fowler:

But surely you also agree that there are some things that are common values.

William Howell:

Sure, sure.

Anthony Fowler:

When we find out that the governor did engage in corruption on some massive scale, that should affect everybody, at least directionally, in the same direction. Those are the things that we want the voters to have... We want them to have that information when they make their decisions.

Wioletta Dziuda:

And if you will, the charter school doesn't have a choice. It is what it is. In politics, politicians have a choice, so presumably the charter school could decide, well, we have more conservatives in a given district, we actually want to have prayer in our school. In this sense, more information would lead to better choice by the charter school, and here I'm putting choice in quotation marks. In this sense, when you apply this logic to politics, that information should lead politicians to choose actions that benefit a larger group of people. So I think in this sense, I think even though Daniel doesn't show us really the behavior of the politicians changes, I think we can be relatively confident that more people splitting their votes gives more incentives for the politicians to make better decisions.

William Howell:

Have you ever wondered what goes on inside a black hole? Or why time only moves in one direction? Or what is really so weird about quantum mechanics? Then you should listen to Why This Universe. On this podcast, you'll hear about the strangest and most interesting ideas in physics, broken down by physicist Dan Hooper and Shalma Wegsman. If you want to learn about our universe from the quantum to the cosmic, you won't want to miss Why This Universe, part of the University of Chicago podcast network.

Wioletta Dziuda:

I want to address one other thing that's related a little bit to what you're saying. Yeah, I think we agree, in general more information is good in this setting. But in our conversation, we sort of equated more information with more access to media, more local coverage of the local politicians. But we cannot forget the type of coverage that the media are providing nowadays. It seems that the media moved away from the model where you want to be the trusted source of news to a model where you want to be entertaining. You want to keep people's attention with catchy stories. I think there's the still big question, which is completely outside of Daniel's paper, but I think we should think about when we think about the implication of his papers and make the policy prescriptions, are people who have access to this media that covers their local politicians, are they better informed, or is that that they're perhaps swayed by crazy stories, by partisan accusations? Perhaps at the end of the day their voting is less informed than the vote of people who maybe don't have access to the news, but they just look at their lives and they say, "Am I happy under this governor? Has my life improved or not?"

Anthony Fowler:

I mean, he has a pretty straightforward measure of political knowledge about local versus national candidates, and what he shows is that those people who live in media markets that don't cross borders, and therefore they get more information about state local officials, they are more likely to know about them. Just basic facts.

Wioletta Dziuda:

Yeah, but what do I know about them? I know their name because I've heard this crazy story about the trip that they took to Cancun or things like that. But perhaps this is not actually the story that I should be focusing on if I want to make an informed decision. While me living in northern Indiana, I might be thinking, "I don't even recall the name of my governor, but that doesn't matter because the moment I go to vote, I know who is the incumbent..." I'm going to just figure out who's the incumbent and I'm going to say, "My life has improved in the last four years, so I'm going to vote for the incumbent." So I think the story that access to media was a good thing is definitely the story of the past, but is this the story of the present? I'm not so sure.

Anthony Fowler:

I think you're raising a very good point. If you do watch local news, I think any particular day, you're not going to get much information that's going to likely be pivotal for your vote. So all of that's fair. Although to say that voters are worse with media would be to not have a lot of faith in their reasoning ability.

Wioletta Dziuda:

Yeah. I'm not saying about you don't change your vote, I'm actually saying you might change your vote based on the information you get. But if you get slanted information, there are a lot of models with reasonable rational voters where you can manipulate voters by selectively giving them information, and you can also think of stories of attention. I'm getting all this information about some strange disputes that politicians engaged in, and I'm basing my votes based on their behavior in those disputes. But then I'm not paying attention to actually what these politicians were doing, what kind of policies they passed and what kind of ordinances they proposed. So I can be rational, but if I—

Anthony Fowler:

Yeah, I get it. But you still have that information, even if you watch the news and you get the slant and all that, you still also get to see did my paycheck go up or down this year? Did the road seem to be better paved or worse paved? So everybody has that information.

Wioletta Dziuda:

True. So maybe I have to confess here that I do believe that there is some behavior component in people's behavior. Like you have... Let me—

William Howell:

I think we have rift on our podcast. For the first time, we have a rift between... Because Anthony in every turn is saying, "But voters will sort it out. They'll make sense of it. They'll have this basic level of information and more information at the margin is only going to improve..." And what you're saying is, "Well, maybe it leads to these distortions," and that the voters in an important way are made worse off, that is they overweigh things relative to what they should be weighting them when they think about in an objective way their own wellbeing.

Wioletta Dziuda:

Yeah. Yeah. I don't know, again, in my private life, I fall behind Anthony very strongly. All my research assumes that people are rational. But I think... I've been watching a lot of TV recently, and I'm starting to doubt my rationality. Before we started this podcast, we had this conversation, Anthony, that you go, you read The New York Times, you form one opinion, and then you go and listen to Fox News and you hear a completely different set of information, and then you form your opinion based on what you've heard from those two sources...

Anthony Fowler:

Yeah. I mean, to clarify, I mean, when you get hyper-polarized news like that, you're essentially getting almost no information. But I think people can sort that out. Maybe not everybody. I mean, I'd be interested to see a compelling study of this, to see maybe people can be bamboozled, but I think there might be a lot of people who see hyper-partisan news and they just say, "Okay. So I guess I didn't learn very much about whether the governor's doing a good job, because The New York Times says one thing and the Wall Street Journal says a complete opposite."

Wioletta Dziuda:

Yeah, but you are a person who goes and reads The New York Times and Wall Street Journal, but think about a person who has always been reading The New York Times, and it had been always providing good information to this person, and then over time The New York Times became less reliable, but it will take you a while to figure it out, unless you go and start looking at other sources of news and some independent journalism. So at least for a while you can trust biased media, and that's going to affect your decision in a negative way. We have research showing that this might be the case.

Anthony Fowler:

It's possible, yeah. I'd like to see the compelling study on that one way or the other. But in the meantime, I'd happily give the benefit of the doubt to the voters and say that more information for them is at least on net in expectation going to improve their decision making rather than...

William Howell:

You mention, what was it, Cancun-gate? Imagine you're in... Is what you have in mind that you are in Texas and you watch a thousand stories on how awful it is that Ted Cruz did this? What that has the effect of doing is up-weighting that behavior in your evaluation of Cruz, and ignoring or down weighting or not being sensitive enough to your lack of knowledge about all the votes that he cast on deregulating this and spending on that. That's the distortion that's introduced.

Wioletta Dziuda:

Exactly. So if I'm actually inclined to like Ted Cruz and I hear the news telling me about Cancun-gate all the time, I can infer, if I don't take into account that the media are trying to manipulate me, I can infer that perhaps there's nothing else that they can say about Ted Cruz, no other negative coverage of Ted Cruz, this is the worst they were able to find, so I'm actually updating my beliefs about Ted Cruz positively. You can put it the other direction, like I'm predisposed not to like Ted Cruz, I think he actually has a horrible voting record and he's horrible for Texas, and now they are on top of that, throwing this Cancun thing, I'm assuming he's horrible, there's nothing good to learn about him, while there might be a lot of legitimate information that they could have provided to me that would actually change my belief. Perhaps he actually is great for the state of Texas. Perhaps he actually delivered a lot on infrastructure in one way or another.

So I'm acknowledging what Anthony's saying, that the voter is completely rational, and they know what's the universe of information that's out there and what's the slant of the media, then they can undo all this biases, but not completely because I really don't know what this information that they're not showing me is. But I think we can believe that, at least in the short term, people can be swayed by this kind of selective coverage.

Anthony Fowler:

I think we're agreeing. I think we're agreeing on that part, that people are swayed by the news they consume, and people can be swayed by campaigns, although I doubt that many people were swayed by Cancun-gate. But in principle, you could be. That does give some power to your local news team, because they can, just by changing what they emphasize, they can potentially change your vote. But that doesn't mean that the voters are worse off than if they had no information. So maybe that's where we have the disagreement.

William Howell:

It doesn't mean they're not, either. We just can't speak to that.

Wioletta Dziuda:

Yeah. I think I'm willing to give Anthony a win, that yes, probably on average we're not harming the society too much by exposing them to local media, but I'm just worried that... You discussed with Daniel possible solutions to the problem, that maybe we should change the media markets, give some incentives to media to increase local coverage and so on, and I'm just saying, given what I see what's happening in the media, how they are chasing entertainment and not necessarily hardcore important information, I don't know how big of a effect of this kind of interventions would be. I'm compromising a little bit, going to the middle.

Anthony Fowler:

Let me compromise too. I'll try to be generous.

William Howell:

Oh, this is no fun. We finally had a cleavage and you guys are like coming right back together.

Anthony Fowler:

I'll reopen it here in a second. But one thing I could say is, one reason why I really don't like a lot of the discussion of the irrational voters is because very often it goes hand in hand with a particular form of paternalism. It goes hand in hand with, "The voters are irrational. They're dumb. And I know what's better for them than they do." American political behavior literature has a lot of that. That makes me very uncomfortable. Even if I can see that, of course, nobody's perfectly rational and the voters are not like solving... They're not using Bayes rule when they watch the news exactly. Nevertheless, whatever way in which they deviate from rationality, I am not comfortable saying, "Here's what will be better for them. Instead of them getting that information, what I'm going to do is I'm going to turn off their TV channels so they can't see that, because I know that they'll actually make better..." That makes me extremely uncomfortable.

Wioletta Dziuda:

Yeah, but I think here we agree. In a sense what I'm saying is, look, I like having TV on and getting some news while I'm cooking my dinner, and I know that listening to CNN makes me actually probably... In one dimension I'm more informed, but in another dimension I'm becoming more biased. I can't explain what the model of me is, but I feel this is what's happening when I'm listening to CNN. So I wish I had actually access to a much better source of information.

Anthony Fowler:

I agree with that completely.

Wioletta Dziuda:

So in this sense, I'm a rational irrational person.

William Howell:

So, Wioletta, we've gone round and round on this, what's your bottom line?

Wioletta Dziuda:

I like the paper a lot. I think it gives a compelling story for why we see people voting for the same party at the national and local level. It seems that voters are rational, and when they do so, it's not necessarily because they just blindly follow the party cues, but perhaps because they don't have enough information not to. I'm left a little bit worried because I don't know what we can do about this. If we think that this voting that we observe leads to lower accountability, I really don't know how we can improve that.

William Howell:

You feel that because there are strong market forces that push in the other direction? Because there are powerful reasons for local media to behave the way that it does, and it's not clear what a corrective would look like?

Wioletta Dziuda:

Yeah. At least that's what we've seen in the last few years. That's not to say we shouldn't be trying. I think maybe there's room for some non-profits to jump in. Perhaps there's room for the government to regulate. But I don't see it clearly. I don't see a very clear policy solution

William Howell:

What do you think, Anthony?

Anthony Fowler:

Well, I share Wiola's enthusiasm for the paper. I think it's a great paper. I think it documents something that is really important in American politics, that in fact more information does change people's votes. On the normative questions, my own inclination is that more information changes their votes for the better, but regardless, I think that's a valuable thing to have documented. I think it's also important because of these big trends that we're seeing with the increases in nationalization and the decreases in local news coverage. I think all of that's really important for us to understand. It changes how we think about how voters behave, what the incentives are for elected officials.

I think a lot of it is normatively reassuring, but as Wiola said, it's not easy to think of solutions to this kind of decline in local news. It's not easy to come up with good policy solutions. Maybe we can have many future episodes about different policy solutions. But on that, I think whatever we could do to try to at least get more information to the voters so that they, when the governor is doing a good or a bad job, they find out about it, I think is probably going to be good on that for a democracy and electoral accountability.

William Howell:

I come out similarly. I think it's a terrific paper. One, it's super careful and clear about what it's estimating. The findings are compelling. And on the horizon are big questions and big concerns about long-standing trends in the media and long-standing trends in the nationalization of our politics, and deep-seated concerns about accountability.

I guess I would say what I hope to see future papers on are efforts to bring some of these concerns that we have into view, that is to see how changes in information induce changes in the behavior of people who actually occupy positions of power, be it by reference to the effort that they give or to the positions that they take. And when we think about the health of our democracy, those strike me as absolutely first order concerns. We as a discipline, political science is doing work on American politics, have got accustomed to saying, "More split-ticket voting, that's..." We see increases in that, we find that normatively reassuring. But that's because split-ticket voting is standing in for all kinds of ideas that we have about what's going on in the heads of voters, and to get more direct evidence of those things would be good.

Anthony Fowler:

Thanks for listening to Not Another Politics Podcast.

Wioletta Dziuda:

Our show is a podcast from the Harris School of Public Policy, and is produced by Matt Hodapp. Thanks for listening.