 Hello, everyone. Welcome. I'm sorry, it's raining for heaven's sakes. So maybe we'll have more people on Zoom today. A couple things I wanted to mention. We have new leadership for Tripoli. This past summer, Mary McEwen was elected president by the board, and then she developed some health issues and had to step down. So this past Monday, the board re-elected me. So I'm still here. And Cindy Macken is the vice president. Where's Cindy? Oh, she's way in the back, in the dark. So she's vice president. Everybody else is still the same. A couple of things. You're going to be getting feedback forms via email, hopefully this weekend. We would love it if you would fill those out and send those back. Same format we've always used. We are planning programs already for the spring. So we would love to see your feedback. And there'll also be an email coming out a couple of events that we're planning for the fall or in winter by the end of December. So look for that on your email. So now I'd like to introduce Matt. So he knows my daughter at channel three. So I better shape up here. Matt Dickinson, now the way we found him is that a lot of you have heard Eric Davis at these lectures. And I called Eric and asked him if he would do it. He said, I'm not doing that anymore. I said, oh, I said, well, whom would you recommend? And guess, guess who? So that's how we found Matt. So the topic of his, he's going to speak the coming red tide forecasting the 2022 midterms. A little about him. He's professor of political science at Middlebury College. He received his PhD in government in 1992 from Harvard University, where he also taught as a professor until 2000. He has a blog. I won't give you the email, I mean the address now, but if you want it, see me afterwards and I will write it down for you. He hosts a weekly politics luncheon open to Vermont residents. He is author of Bitter Harvest, FDR, Presidential Power and the Growth of the Presidential Branch. That was in 1999. He's the co-editor of Guardian of the Presidency, the legacy of Richard E. Neustadt and has published numerous articles on the presidency, Congress and the executive branch. His current book manuscript titled, Clerk or Leader. The president and the White House staff, people, positions and processes 1945 to 2020 examines the growth of presidential staff in the post-world war two era. So please welcome him at this very opportune time. Matt. Thank you, Carol. Thank you, Cindy, somewhere, and thank you all for showing up. I'll help you know it's sunny in the rest of Vermont. It's literally the only place we're training, which is why we almost didn't find this. As Carol says, I teach at Middlebury College and at this time of year, can you hear me in the back there? Okay. I get trotted out, go around the state and other places to predict the outcome of midterm elections. Every two years they trot me out. I'm going to do that today. But I should warn you, my prediction is based on a lot of political science forecasting models. So what I do is sort of put them all together and then present the forecast. And you might ask, should we trust this guy? We've never seen him before. And so I'll just point out, and this is talks that I've made in previous years. The predictions are what I, this is for the house now, which is a lot easier to predict than the Senate. On the left hand is what I predicted, the number of seats and the actual house results. So, you know, that's not bad for government work. Of course, like stocks, past performance is no guarantee of future results. So with that caveat, let's plunge in. And let me also say, I'm going to try to do a hard stop here at about quarter three, because Carol says you guys get kicked out. They turn off the lights at three or something close that I would love to have your feedback and some questions on either this or anything having to do with politics. So I'm going to go through a lot of PowerPoints relatively quickly. If at any point you have a question, just stop me and I will try to clarify it. I also tend to roam on stage quite a bit. This isn't quite a stage and I see there's a camera on me. So I'm going to try to glue my hands here. But if the podium begins moving across, you'll know what's happening here. And again, I also try to predict the Senate. And the Senate is a lot harder to predict. There's fewer Senate seats up. You might think that makes it easier, but it actually is harder because they're much more idiosyncratic. Senate races are more competitive. Outcomes have much more to do with the candidates. The Senate voting public is much more heterogeneous. So I've done pretty well with the Senate, but this year is going to be, it's quite, it's quite a circus. All right. So the talk basically, I'll talk about the electoral context, what's at stake. Then I'll present something that comes as close to a law of political science that we have. That's the midterm loss phenomenon. I'll talk about a few developments like the abortion decision that might throw some of these forecasts off. Then I'll give you my forecast. Actually, I think I'll give you the forecast first. And then very briefly talk about the implications if these forecasts are accurate for Joe Biden's ability to govern going forward. So as you all know, every two years, every house seat, all 435 are up for election. Currently, the House Democrats have a slim majority. They occupy 2,219 seats. The Republicans, 2,211. There are some vacancies that make the calculation of how many seats the Republicans must gain to gain a majority somewhat uncertain here. But roughly they have to flip eight seats the other majority. That's not hard to do based on historical record here. The Senate, on the other hand, is tied 50-50. As you know, only a third of the Senate is up for election every two years. So we have 34 seats this year up for election Democrats defending 14 Senate seats, Republicans defending 20. This is the House map, 219 Democrats in blue, 2,211 Republicans in red. Of course, it looks like the Republicans dominate because our house seats must have the equal population. And in many of the democratically controlled seats, the population is quite densely located. Think about all those urban areas on the east and left, the east and the left coast, the west coast. And so that's why it looks like geographically the Republicans hold more territory they do. But population-wise, the Republicans are actually a minority. But this is going to come into play, this geographic distribution for reasons that I'll talk about in a moment. So remember this map. In the Senate, of course, the white Senate states, there are no Senate races up. The blue states are currently held by the Democrats, 14, the red 20 Republicans. And of course, again, that's a 50-50 map here. So any single state here is crucial to the outcome of control. But when we're talking about competitive Senate seats, there are really only about seven, or perhaps six that are competitive. Of those, the four, Arizona, Georgia, Nevada, Wisconsin, all currently have incumbents running for reelection. That's important because incumbents generally win reelection in Senate races about 85% of the time. And then there's a couple seats, Pennsylvania and North Carolina, where the current incumbent is stepping down. Those are open seats. Pennsylvania, possible pickup for Democrats. North Carolina, another possible pickup for Democrats. Then there's a couple of seats that I don't think are really in danger of flipping New Hampshire, Florida and Ohio. Ohio might surprise you. That's an open seat. Tim Ryan running against J.D. Vance. For reasons that I'll talk about, I think J.D. Vance has a pretty strong advantage in Ohio, a state that's been trending Republican. But here's what I really want you to focus on. This is called the midterm loss phenomenon. It is, there's three things in life that are certain death taxes and that the president's party loses seats in the midterm election. As you can see, that green are the three occurrences in which the president's party actually gained seats since 1934. In all the other races, the president's party lost seats. And so this would lead you to believe if you're a betting person that heading into the November elections, the Democrats are going to lose seats and most likely lose enough seats to lose their majority in the house. You might not know, but this phenomenon also affects the Senate. Now the Senate is a little more idiosyncratic. There's only been five instances, however, in which the president's party has gained seats in the midterm election in the Senate. So we have this midterm loss phenomenon that almost always strikes in the house and more often than not in the Senate. And the question is why? What explains this? And we are political scientists. We're not pundits. We profess to understand politics science. That's why we call it political science. And the science requires that theory and explanation and then evidence on whether that theory is right. So there are three theories or explanations for this midterm loss phenomenon. All three may come into play in November or maybe just one. I will briefly run through the intuition behind all of them. The first we call the surge and decline thesis, which points to the difference turnouts in presidential election years versus the midterm. The second is the midterm is a referendum on the president. And for reasons that we will discuss, typically the people who don't like what the president is doing are more likely to vote in the midterm than the people who like what the president is doing. And third, and this is particularly, I think pertinent to the current polarized era, Americans are largely centrist in views, but our elected representatives are much more polarized than our political activist class is much more polarized. And so the Americans in their wisdom don't trust either party to hold a majority for very long. Instead, they tend to balance off control of party. And right now, since the Democrats are in the majority, the balancing thesis tells us if enough Republicans can coordinate and Democrats as well, those who are dissatisfied with the Democratic party, they're going to vote for the Republican candidates to balance off control. So these are our three explanations. Is there evidence for any of them? Let's look at the surge and decline theory. The theory is quite simple. During a presidential election year, let's think 2020, we have a surge of voters, people who don't necessarily pay attention to politics on a daily basis, but they will vote in a presidential election. Note that turnout in the presidential election of eligible voters is roughly 60%. And then this additional surge of voters are people who are very responsive to the politics of the day. But what happens in the midterm 2022? A good chunk of those individuals are simply less interested in voting because once you get away from the presidential contest, they're not quite sure whether they're going to participate in a midterm. They often don't know who their representative is. And so they stay home. And that erases the majority party's advantage because proportionally, the people who stay home are more likely to become from the winning coalition in the presidential election year. So if you look at turnout, that blue line is presidential election year turnout. And if you look at the red line, that's midterm election turnout. You can see there's about a 20% difference between the two pretty consistently. So the surge in the presidential election year, that 60%, leads to a decline of turnout in the midterm, about 40%. And most of that decline is among the president's party. Therefore, in the midterm, president's party tends to lose support. How much? One way we estimate this is by looking at independence. Why independence? Independence are the individuals who are less committed to either party. They provide the change from election cycle to election cycle. They are growing in number in the United States. I'll talk about that a bit later. So they're much more responsive to the mood of the moment. They're the ones who give the president the winning candidate in the presidential election his margin of victory, usually. So in 2020, 94% of Republicans voted for Trump. 94% of Democrats voted for Biden. So what made the difference? It was the independence. Biden won them by about 51 to 43%, something like that. This is why it matters. We know for each percentage point advantage that the president had in the presidential election among independence alone since 1980, that predicts a loss of 2.2 house seats on average in the midterm. Because the independence don't vote in the midterm. And so all we have to do is look at the advantage Joe Biden had among independence in 2020 and multiply that by 2.2 and that'll give us our first ballpark estimate of the number of seats the Democrats are going to lose. So he won, according to Thomas Patterson from the Kennedy School of Government, a 14 point advantage among independence in 2020. Multiply by 2.2, that predicts a Republican gain of 31 seats this November, more than enough to regain the house. Remember, only need about eight to 11 seats to regain the house. Second theory, referendum theory. This is the one that says the midterm is a referendum on the president's performance in the previous two years. After taking office, presidents make a lot of choices. Joe Biden has made choices in infrastructure bill. If you go down the list of the negotiations he's had with Joe Manchin and Kirsten Sinema within his own party, he hasn't been very successful getting Republicans to support him. And these choices tend to produce more enemies than friends. Your people who are most passionate in the midterms tend to be people who are disappointed with what happened. And unfortunately for Joe Biden, his approval ratings, this is from real clear politics. So it's an aggregate. Each of those lines is an aggregate of multiple polls at any particular time. The red line is the number of Americans who disapprove of his performance. The black line is the number of Americans who approve of his performance. You can see until really the beginning of the fall, August beginning, September, his approval ratings had been steadily declining, his disapproval ratings steadily increasing. They've rebounded somewhat, and you can probably figure out a couple of the precipitating factors here. One, of course, is the passage of key legislation, the anti-inflation bill, which is inflationary, but never mind that. And the Supreme Court decision on abortion has given a temporary boost, and I emphasize temporary, to the Democrats. And so his approval ratings have come up a little bit. He's now at about 43 percent approval. That's about two percentage points higher than Donald Trump was at a similar point in his presidency. And it points to Democrats losing seats in this referendum. How much? Well, if you draw a line based on past elections dating back to 1980, that blue line is a regression line. It basically says what's the relationship between the president's popularity and the share of the other party's popular vote in the House now. We're just looking at the House. You can see if we go over to about where Biden's popularity is on the bottom, that X horizontal line, let's say he's at 43, 42 percentage points approval, you go up on that line and you look at the Y axis, the vertical axis there running from 46 to 56, it says the Republicans, the out party, are probably going to win about 53 to 54 percent of the popular vote. That is all the votes cast for all the 435 House seats. Now you say why does that matter? We don't elect the House based on the popular vote, but we can do a rough translation from the House popular vote to the number of seats they're going to win. Now you might say, well, if you win 53 percent of the popular votes, you're going to win 53 percent of the House seats, but no, you're going to win more than that. And that's because of the peculiarities of the American political system, it's winner take all. So unlike hand grades or bowling, you don't get points for being a close second or horseshoes. If you're a close second in a House race, you get nothing. You don't win the House seat. It always goes to the person who gets the most popular votes in that district. If you add that all up over time, it leads to a bias in which the party's candidates who are winning the most popular votes are actually going to win a greater proportion of seats in the House. How much greater? Well, if you look at this chart, these jagged lines represent the sheet, the seats one in each election over the proportion of popular vote. So if there was zero, if it was exactly the same, there was no difference between the proportion of seats and the proportion of your popular vote in the aggregate. It would be down at the bottom as zero, but it's never down at the bottom. It's always higher. That is, you always win more seats proportionally than your party wins popular votes. How much more on average between four and six percent more seats than your popular vote? So again, extrapolating from history, if you're going to win 53 percent of the popular vote, you're going to win between 57 and close to 60 percent of the seats in the House. So just to do the math very carefully, keep in mind now, I pointed out this map earlier, and this is where this comes into play as well, the Republicans are more efficiently distributed. That is, Republicans hold more seats in which their margin of victory is quite small over the Democrats, whereas Democrats hold more seats in which their margin of victory is huge over the Republicans. Think about a place like where AOC represents New York. She won her last election with about 85 percent of the vote. There's no Republicans in New York City, for the most part, but a lot of Republicans are winning in areas in which it's pretty evenly divided. What does this mean? It means Democrats are wasting votes, a lot of their popular vote because you don't get extra seats within a district just because you won 80 to 20 as opposed to 51-49. So two factors are working against Democrats here. One is what I talked about earlier, this structural result of the single-member simple plurality voting system, and two, the inefficiency with which Democrats are distributed. What does that mean? For example, in 2014, the Democrats won a House seat for every 189,000 votes they won in a district. Districts right now are roughly 700,000 people. The Republicans, by contrast, only had to win 162,000 votes to win the seat, so that is a structural advantage favoring Republicans. They win more seats based on the same amount of popular votes. In other words, if as many people vote in 2022 as did in 2014, the Democrats need to win about 5.3 million more votes than the Republicans just to break even. And they're not going to do that, we don't think. So let's make it very simple. If the Republicans win 53% of the popular vote, adding up all the votes in all 435 House races, 53% of it goes through Republicans, they're going to pick up based on past performance about 35 seats, according to this model. So we've got two forecasts here based on slightly different theories, but we're not done. There's a third theory, we call this the balancing theory. Now, this theory is not going to give me an exact number of seats of prediction. And why do we balance? Well, there's different explanations. We're balancing because we think power corrupts in absolute power, corrupts absolutely. So the worst thing you can have is one party in control of the House, the Senate and the presidency. We also think as both parties have become increasingly internally homogenous, ideologically speaking, Democrats are all really tree hugging eco-terrorist, pre-estriving secular humanists, right? And the Republicans are increasingly Bible thumping, gun-toting, neo-Nazi goose-stepping. You get the point. Normal people like you and I can't stand either of these parties. And so we're just going to balance them off. And then there's an institutional explanation that says when we vote for the presidency, we're looking for different qualities than when we vote for Congress. And so we vote differently in the presidential election year, where the top of the ticket drives our vote than we do in the midterm election year. All of these explanations may come into play. The point is we expect a balance coming up in 2022. And with the Democrats holding majorities, the balance is going to favor the Republicans in some form. And just to give you a sense of how polarized we are today, this red line represents the House. This blue line represents the Senate. It goes back to the 1880s. That is the post-Civil War reconstruction. And what this, these lines are measuring is that if you basically look at each member of the Senate and the House's voting record, and you assign them an ideological score. So Bernie Sanders score is negative 12. He's somewhere left of Mao Taisung, right? And you look at somebody like Rand Paul, who's somewhere right of Genghis Khan. And you array everybody there, and then you just take the median member, okay, where's the median score for all the Democrats in the House and the Senate and the median for all the Republicans. And you say, how far apart are those medians? And that'll give us a measure of how polarized the chambers are. And graph it, that's what we have here. The higher the point, the more polarized the two parties are. Look at us now. We are more polarized in the Senate and the House than we have been since the Civil War. That's how bad things are right now in terms of polarization in our Congress. And so that balancing calculus really comes into play here, because the public is nowhere near either of these parties in general. Now, the political activists are, but Jane and Joe Sixpack, normal people who don't really live and die by what they see on cable television is we're much more moderate. We have a mix of views, some conservative, some liberal, but we're not purists by any sense of the word. And one of the things we know when we look at polling about how voters say they're going to vote heading into the midterm. And if you track that polling in the weeks before the midterm, it's the out party that gains support. The out party tends to become increasingly more popular. And that's because of this balancing, as people begin to pay attention to the race and they say, who's in charge now? I think I'm going to vote for the other party. Not everybody does that, but enough to sort of lead to this balancing. We don't know how that translates into seats, but let me just show you this. If you think about the majorities in the House, in the Senate, in which party controls the White House, you really can only have eight configurations. You can have all Democrats, you can have all Republicans, you can have some mix of Republican president and a divided House and Senate or Democratic president and Republican House and Senate. There's eight possible permutations. Well, since 2001, the beginning of this century, we've cycled through every single permutation except one. The only one we haven't had is a Democratic president, a Republican Senate in a Democratic House. And we came within a whisker of doing that in 2020. If Donald Trump had not scared the Republican voters in Georgia in the special election by saying your vote isn't going to count, and we know in doing that a lot of Republicans didn't vote in those special elections, those Senate seats would have gone Republican in Georgia and we would have had a Republican-controlled Senate. In other words, Americans aren't satisfied with any configuration of control since really the beginning of this century. And they just keep cycling through saying, whatever you've given us after two years, we don't like it. And we're trying something else. And so if you have a Democratic majority right now, we can't get rid of Joe Biden. I don't think. And so we're going to have to change control of the House and the Senate. So this leads again to the prediction. Since the disputed 2000 election, we've experienced seven of the eight possible configurations of partisan control. No single configuration has lasted a full presidential term. We essentially are being governed by minority parties. There is no majority party. We've had no landslide presidential election since Ronald Reagan. We are in a period in which the majority of Americans do not like either party enough to give them control. It's just the nature of where we are. And an additional measure of that, when we ask people, do you identify as a Republican, a Democrat, or independent, the percentage of people saying independence now is the largest it's been since we've been asking that question. It's over 40%. So it's all reasons to think voters are going to reject the Democrats this fall in some form or fashion due to this balancing theory. All right. Let's summarize what we got here. Surgeon decline says pick up 35 seats for Republicans. Referendum on Biden 32% seats for Republicans. Balancing model, no clear seat prediction, but clearly the Democrats will lose their House majority and perhaps the Senate majority. I'm going to get to the Senate in a moment. Now, there's a lot more forecast models out there. Political scientists love their forecast models. I'm not going to run you through more of them, a couple more based on looking at presidential approval. So that would be the referendum model plus disposable income. It's a measure of how Americans are doing in terms of the economy. That predicts the Democrats are going to lose 44 seats. Another model which looks at the number of seats that are competitive and based on past history looking at the number of seats that are competitive. The prediction is Democrats are going to lose 42 seats. So all of these models are forecasting a loss of the Democrats seats from anywhere to 31, what was it exactly there? 32 all the way up to 44. If, for instance, the worst scenario for Democrats come true and Republicans pick up 44 seats in the House, they will have the biggest majority in the House that they've ever had in history. And of course, if you're a Democrat, that's not good news. And so you immediately say, what happens in the Senate? Please Lord, let us hold the Senate. But before we get there, I want to give you a dose of good news if you are a Democrat. And I don't presume you are all Democrats. There is a voice out there, a voice in the wilderness of political scientists who says, wait a minute, there's another model here. And this model suggests the Democrats might be saved. And they might be saved because history is not going to work exactly as it has in this fall because additional factors are coming into play here. And the most important one is we're losing competitive seats, seats in which it's actually possible for the out candidates to win. The incumbency factor has always been strong in House races. When incumbents run for re-election, they win about 95% of the time in the House. But they may even win at a higher rate in this cycle. And the reason is because there are fewer and fewer seats that are considered competitive. And I'll talk about why that is. But when you think about landslide, or as I called it, wave elections, the red tide, in the past, they usually took place when there were many more competitive seats. And usually because one party had a huge majority. And then then their majority was wiped out. In this current environment, neither party has a huge majority. So for, excuse me, in 2010, when the Obama and Democrats underwent that shellacking, do you remember that? In 2010, Democrats lost 54 seats. Well, they had a 79 seat advantage at that time in the House. And so it enabled the GOP. Actually, that number is wrong. It was less than that. But the point is they had a huge majority. They do not have a huge majority now. And Alan Mbramowicz, a political scientist down in Emory, has actually looked at the history and he's created his own forecast model. And that model is based on two criteria. One is how big an advantage does the majority party have? So the Democratic Party now has an advantage of roughly 11 seats. And then the second measure is we go out and we ask voters, are you going to vote in the House election? And if so, will you vote for the Democratic candidate or the Republican candidate? And we just call that something called the generic ballot question, because we don't ask them the candidate's names, because most voters don't know who the candidates are. They just know I'm going to vote Democrat or Republican. And if you factor that in and look at past performance, you can predict based just on the current generic ballot and knowing the Democrats have 11 seat majority, what the outcome is going to be this November. And so these are all the past predictions and the actual results. That line would be the predicted seat change in previous House elections based on this model from 1946 to 2018. And you can see it more or less does a pretty good job predicting where the actual outcome is. So what does that model predict here? Well, here's the generic ballot. You can see it's been fluctuating, but I just looked this up today. So when we ask a survey of Americans, are you going to vote for the Republican candidate for the House or the Democrat? A slight, slight plurality, say the Republican, it's less than a percent. So it's really close, the generic ballot. And if we plug that into his model, the difference between the Democratic percentage and the Republican percentage that's in that first column will lead to a prediction of how many seats the Democrats will lose. So we're just between zero and negative five in the generic ballot. We're closer to zero. What does that predict? Yeah, really just only a loss of maybe 15 to 20 seats. So this model still has the House flipping, but by a lot less. And that should give you pause before you go slit your throats as a Democrat and think you're going to lose the House majority and it's going to be a title wave. It might just be a trickle, but you're still going to lose it. Okay, is there anything that could upset these predictions very quickly? These are the issues that people usually talk about, the pundits. I'm not going to spend a lot of time here because I want to make sure I leave you time for questions. Let me just briefly remind you this is 2022. So it's the first election after we redrew congressional lines. You know, we do that every 10 years. US Census is taken. The population in each House district might shift. We have to redraw those district lines. That is a political process. It's done by the state legislatures or neutral redistricting. Parties actively try to use this process, which we call gerrymandering, to ensure their own members will win reelection and the other parties members won't. This is the first election since that happened, so that may make some of our forecasts slightly off. I'll talk about that in a moment. Generally speaking, if we estimate the impact of gerrymandering, the redrawing of lines after 2020 census, we think the districts that went Democratic voted for Joe Biden, we think Democrats have picked up one more. They're now at 188 districts that lean Democrat. The Republicans picked up 10 more as a result of redrawing lines. There's probably 168 districts that are now strong Trump. Most importantly, the number of competitive districts has declined by 11. This is a long-term trend. Of course, the definition of competitive varies. Based on this measure, the number of competitive seats, one defined as one in which the presidential candidate in the prior election were within 5 percentage points in that district. We only have about 40 competitive districts in 2022. And again, that's the Ambrama Woods model. So gerrymandering may limit the scope of the Republican wave here because a lot of districts, it doesn't matter how strong that wave is. They're just not competitive. The incumbent or the incumbents party, just because of the way that the lines have been drawn, there's just so many partisans in that district. It's going to be very difficult for the opposition party to take that. One way to think about that is just look at the map. And those gray, or I don't know what color that is, in the middle there, those are the districts that are viewed as pure toss-ups. There's just not a lot of them, if you can see in the house here. It's also the case, of course. We want to look at, we know partisanship is the most important predictor. So what's the relative proportion of individuals calling themselves Republicans or Democrats? Throughout 2021, the percentage calling themselves Republicans was trending up. Republicans, sorry, Democrats trending down. But note, the biggest gainer, independence. That's the current breakdown as of July. I haven't been able to update this. 41% say they are independent, 29% Democrats, 28% Republicans. When you force people to choose a party, they say, okay, I'm a Republican or a Democrat. But if you give them the choice of, or maybe you're an independent, a lot of Democrat and Republican leaners say, no, I'm actually an independent. Why does that matter? They're less likely to vote in the midterm than they were in the presidential election year. Okay. The other thing we look at is enthusiasm. We have been looking here at seat changes as a function of these factors, but turnout matters. You got to get your people to the polls. And one way we predict turnout is looking at enthusiasm. And right now, the enthusiasm for voting in the midterms is pretty close. I wasn't able to get an accurate number as of October. So I'm going to give you a summer number. But if you look in previous years, you can see, for instance, in 2010, that was the shellacking in which the Republicans gained 60-some seats. You can see the enthusiasm gap was huge. Similarly, in 2018, the Democrats had a slight advantage in the enthusiasm gap. They happened to regain a house that year. So the enthusiasm gap is pretty much a good predictor. If you're looking past years, when there's been a huge gap in enthusiasm, that's when we most likely to see a shellacking. So on the right-hand side is the enthusiasm gap, which party it favors. The left-hand side on the bar here is the change in seats. There's a rough correlation. Where's the enthusiasm gap today? Eh, favors the Republicans by about 3%, but I have seen more recent polls that suggest that it's tightened even more. So not a huge advantage for either party here in terms of the enthusiasm gap. I haven't said very much about issues. Don't issues matter. Yes, issues matter. Most importantly, we know the economy. Biggest impact on, in fact, I showed you a forecast model that looked at the change in disposable income as one of its influences. I haven't said much about the economy because, to a certain extent, Joe Biden's approval ratings reflect the state of the economy. The more you think the economy is going bad, the lower his approval ratings are. So just by measuring his approval ratings, we're kind of taking into account the economy. There are other issues, of course, most notably the abortion issue. That, ruling by the Supreme Court, gave a slight boost for Democrats, but it is waning right now. The impact of that is receding as I speak, as other issues come into play here. So let's look at the economic measures. I don't have to tell you that inflation is at the highest since it's been 1981. It's not going to help the Democrats. The consumer confidence is dropping. It's dropping to levels not seen since the last recession. There are indications that we are heading into a potential economic recession. Again, not helping Democrats. If we think about members who are stepping down, members are very strategic. When members decide not to run for reelection, it's usually because they think they're going to lose, possibly because the national circumstances are against them. Well, who's stepping down? 22 Democrats compared to 10 Republicans have decided not to run for reelection. Another indication that they can see this upcoming tide coming in. So we might think that abortion is going to rescue the Democrats, and we know the number of individuals who favor legal abortion for any reason is gone up, and the Supreme Court decision is not very popular because of that support for abortion rights. But if you drill down more deeply, Americans are quite nuanced in their views on abortion. If you ask Americans, are you pro-choice or pro-life, a slight majority will say we're pro-choice, and that's pretty consistent. But most Americans are not purely pro-choice or pro-life at all. They are right in the middle. They want abortion to be safe and legal up to a certain point. And that point is somewhere around fetal liability, and then they are in favor of restrictions on abortion rights. And so the idea of the Supreme Court throwing that decision back to the states, it really matters which state you're in about whether that decision is going to have political consequences or not. And so I think the media has been overstating the impact of the abortion issue on the midterms. I think it's going to have much smaller impact than you might think, and not always in the direction you might think. The other thing to keep in mind is we often think about abortion in terms, it's a woman's issue, and it's driving women to the polling booth, and they're going to vote Democratic. In fact, when you ask individuals whether they are supportive of abortion rights or not, historically men are slightly more supportive than women. Not by a lot now. I don't want to overstate this. And it sometimes varies. Women are slightly more. But the point is in terms of its political impact, this isn't quote unquote a woman's issue in terms of how it's going to affect the vote, since so many men about equal to the number of women are quote unquote pro-choice. So when you look a lot of these stories about women, the boost in registration, that's not telling you the whole story about the impact of abortion. So just the caveat there. All right. Well, that's about it. Let me do the Senate very quickly. Bear with me. Again, this is the Senate map. We know 20 Republican-controlled Senate seats are up, 14 Democrats. That generic ballot model, you can apply that to the Senate. And it says the Democrats are probably going to gain a seat or maybe it's no change at all. So we're looking very close in the Senate. So let's look at some individual races. And again, remember, just like in the House, the Senate races have become increasingly non-competitive. This is the number of Senate races in a particular state that trend the same way as the presidential election in that state. And you can see as the state goes in the presidential election, so it goes in the Senate race. So we're not going to see a lot of change in the midterm here. When you think about the Senate delegations that are split, that's the gray one at the top. We're at the lowest number of split Senate delegations we've ever had. All a long way of saying there's not a lot of Senate races that are competitive. If you are a Democrat, however, you have to be pleased with the set of characters the Republicans have put up in some of these Senate seats. And again, I'm not in Georgia. I'm not in Pennsylvania. I'm not in Arizona. That's Brookmaster, Blakemaster, sorry, at the top, Herschel Walker, bottom right, Mehmet Oz. I don't mean to disparage their candidacies, but they're an odd group, and they're a controversial group. And it's unclear to me that these are the strongest candidates that the Republicans could have put up. So you do as a Democrat hope that the Republicans are snatching the feet from the jaws of victory here by running these candidates. Having said that, I'm going to tell you as I make my forecast, my forecast has changed since the last speech I gave on this at some other place. And the reason it's changed is because see these characters, they're all actually tightening the race. All three of these Republicans are actually gaining ground on the Democrats. I didn't think that was going to happen, but it is happening. It's not clear they're gaining enough to win, but they are making these races a lot closer. And that's consistent with what I told you. As you go from the beginning of an election year towards November, the out party tends to gain support. It's happening in Pennsylvania. It's happening in Arizona. It's happening in Georgia right now. Will they gain enough to flip it? Hard to tell. All right. So I'm going to give you my forecast. And I'm going to warn normally my forecast, I just averaged together the political science models and say that's my forecast. I'm cheating this year. And so those of you who came here and you've got your grandson's tuition money and you're going to bet on the outcome of this, don't do it. I am giving the Republicans a gain of 28 seats, which is lower than most of the models are predicting and I think is probably lower than what's going to happen. Again, I made this prediction about a month ago, so it's now public record. Since I made it, I think actually the Republicans chances have gone up to gain a few more seats. The Senate, I'm thinking it's going to be a 50-50 split here. The Republicans will win five of these toss-up states, but of those five, only one will be a pickup. That'll be Nevada, where Adam Laxalt is the son of the Laxalt family is pulling ahead of the incumbent there, Mastro. The Democrats, I think, will pick up four, but only one will be flipped and that'll be Pennsylvania. I'm looking at Federman holding on in Pennsylvania over Mehmet Oz. That means the Senate will remain 50-50 and assuming they can hold together, that will give Joe Biden a working advantage. And again, if you're a Democrat and you have a choice, do I want a majority in the House or the Senate, you want the Senate. That gives you control over court appointments, for instance, nominations. Assume you can hold that together. Forget that. Consequences. Let's assume Biden loses his House majority and potentially the Senate. It matters. We saw it with Barack Obama when he lost his majority. We saw it with Donald Trump when he lost his majority. We saw it with George W. Bush when he lost his majority. We saw it with Bill Clinton when he lost his majority. In each case, their ability to pass legislation declined precipitously and they had to rely on administrative. So here is the percentage of bills passed by Congress backed by the president. On the top, when the president's party is in the majority, 83% of the bills are passed. On the bottom, when one or both houses are controlled by the majority, the percentage drops to 56%. This is Trump's legislative success rate from his first two years to his last two years. Dropped from over 90%, relying on Mitch McConnell and Paul Ryan in the House to getting almost nothing through the democratically controlled Congress. Same with Obama. 96% in his first year, 88% in his second, dropped to 56% after the shellacking, the Republican regaining the House in 2010. The only other thing I would tell you if the Democrats lose the House, you're saying, oh, what does this mean for the presidential election? It means nothing. Presidential elections are a whole other beast. You can invite me back and I will give my presidential election forecast. I'm much better at presidential elections. But half the time, the winner of the midterm has won the presidency. Half the time, they have not. Okay. Let me stop there. Open it up for questions. If you have any. Thank you. That was great. Okay. Any questions on Zoom? No questions on Zoom. Come on, people. If you haven't been zooming before. So if they have, if you haven't, don't have any questions on Zoom, please think of some. Oh, good. We have some live people. Is it just me or has the accuracy of polling declined in the last 10 or so years? It is not just you. The question for those of you who might not have heard it was, has the accuracy of polling declined in the past 10 years? It has. There are a couple of reasons for that. So for instance, in 2016, you'll recall heading into the presidential election, it appeared based on polls that Hillary Clinton was going to win. And when you looked at state level polls in particular, it appeared that she had more than enough for an electoral college majority, but it turned out in several states, the polls understated Donald Trump's support. And the reason it understated, we think it's a combination of several factors. The most important one is the turnout among individuals with a high school education or less. When you poll a state, you're taking a representative sample of who you think the voting public is going to be. Well, almost in all of those states, particularly in the Midwest, Rust Belt states, like Michigan, Wisconsin, that flipped from Barack supporting Obama in 2012 to supporting Clinton, the polls did not estimate the number of lower educated voters that came out and turned out in favor of Trump. And so they were completely wrong. However, if you were smart and you went to one of Dickinson's talk, and he gave you a forecast model, remember our forecast models are not based on polls. That's the great thing about our forecast models. They're based on factors that I would have told you that Hillary Clinton was going to win the popular vote with about 51.5% vote, which by the way, was exactly right. Now, there was that other part of the equation, which as I said, and therefore should be the next president of the United States, which turned out to be slightly wrong, but never mind that. That's not theoretically very interesting, the fact that I predicted she'd be the president and she wasn't. The point is, I understood the popular vote. Now, normally you don't get a split between the popular vote and electoral college vote and we forget we think of 2016 as this huge upset and how could Donald Trump have won, but it really wasn't. He got about the share of the popular vote we thought he'd get. It's just that he managed to, it was as if he drew a straight flush. If you shifted 40,000 votes, which is less than one-tenth of 1% of the total votes and you move them a few miles across state lines, Hillary Clinton would have been president and nobody would have ever heard of Donald Trump ever again. So, one of the oddities of our presidential selection system is in a winner-take-all system in which the winner of the popular vote in each state gets all its electoral college votes, the location of the vote matters a lot. That's a long answer, but your intuition is right. Polling has become more difficult because the composition of voting has been off based on what pollsters thought. A lot of people have cell phones and the random sampling of cell phones wasn't allowed under federal regulations and if you just called people on landlines, you're getting a lot of older people. You're missing a whole bunch of other people. And the other factor is people just aren't answering the phone anymore. I mean, I don't know about you, but before I left this morning, I got 15 calls, random calls from people trying to sell me things. I don't even answer the phone. If you're a pollster, you've got to call more and more people and it gets increasingly expensive and so there's a lot of polls on the cheap. So what pollsters are now doing is they're going online and they're asking people to answer surveys online, but it becomes complicated there because the people who are willing to do it, they're not representative of normal people. They're more politically active, so you've got to balance it off. So yeah, you're right. Polling has become more difficult. I would take polling with a grain of salt heading into 2022. Again, it is not that people are putting their fingers on the polls. Most reputable pollsters are doing the best they can. It's just hard. It's a lot harder. So I have a question and you almost answered it. Do you think the Democrats have waked up since they stayed home and didn't vote for Hillary? Well, so again, there is some debate about whether they stayed home and didn't vote for Hillary. She got roughly 100,000 votes less than Obama. That's a drop in the bucket. In other words, the coalition that supported Obama came out for her. The biggest problem she had was that drop was disproportionately among voters of color. And that meant in some of these swing states, particularly among Latinos, she lost support. And in fact, Joe Biden, heading into the 2020 election, lost more support. Of the four years of Trump, we saw a shift of voters of color from the Democratic Party to the Republican Party. So my worry as a Democrat, it's not total turnout. It's who's turning out. The Democratic Coalition is increasingly becoming wider. It is becoming more affluent. And it's losing working class voters, particularly working class voters of color. Now, if that continues, the Democrats are going to be in trouble. So as Democrats, you got to think about not just the total turnout, but how can I regain some of those working class voters that are that I've been losing? And I think that's the worry Democrats have going ahead. Regarding October surprises, what if Putin actually followed through on his threat to use nuclear weapons in Ukraine? How would that? Great question. What I would say is the question was, what if Putin uses nuclear weapons or something happens in Ukraine? How would that affect the midterms? So you remember that chart where I talked about the exceptions to the midterm loss phenomenon? One was 2002. What happened? Prior to that, 9-11. The other one was 1998. Do you remember what happened then? I got two words for you. Monica and Lowinsky. There was an impeachment of the president. So history suggests if something of that magnitude, an impeachment or a terrorist attack or the use of nuclear weapons, all bets are off. And my guess is there'll be a rally around the flag effect. If again, again, it depends on, you know, what's the impact of those nuclear weapons? Does it affect the United States? And if so, how? But I think that would probably boost the Democratic support. I'm guessing here it depends on the nature of the impact. We have a couple of Zoom questions. What is the influence of Fox News on elections? Not much. Again, people overestimate the number of individuals watching Fox News, Fox News gets maybe three to four million viewers. We had 170 million people voting in the presidential election. The people who watch Fox News, their views are really baked in. It has very little impact in terms of elections. It's about similar to MSNBC, slightly smaller audience, but the people who watch MSNBC are, again, I caricature them, but they're, you know, the tree-hugging eco-terrorists. The people watching Fox News, they're the goose-stepping, yes, you, they're the goose-stepping neo-Nazis. They are activists. And one of the things I caution you, never draw conclusions about the status of American public opinion on politics by watching cable news. But yes, they have on, aren't representative of the views of most Americans. The issues that they talk about are not what most Americans are concerned about. It's an entirely misleading picture. And we tend to overemphasize the importance of those views that are being expressed in terms of understanding American politics. They are completely misleading. They give you the sense that Americans are deeply polarized and they talk past each other and they can't even agree on facts. That's not true. Americans are mostly centrist. They hold a mix of what we might call Republican and Democratic views. And they share much more in common than what divides them. I have a whole nother talk on that issue. I think that we are shutting down, but Matt might stay a little longer for some additional questions. Yes. Thank you, Matt. Yay.