Sabatos Crystal Ball

Introduction: The 2018 Midterm Election Forecasts

James E. Campbell, Guest Columnist September 13th, 2018


Editor’s Note: In an effort to provide as broad a view as possible to readers about different methods of forecasting the 2018 midterm election, we have been featuring models from respected political scientists that aim to project the net seat change in the U.S. House of Representatives. So far, we’ve published models from Crystal Ball Senior Columnist Alan Abramowitz and from the team of Joseph Bafumi, Robert Erikson, and Christopher Wlezien. Both suggest the Democrats are favored to retake the House majority. This week, we’re featuring two more models from James Campbell and the team of Charles Tien and Michael Lewis-Beck. They too forecast a Democratic House takeover, and of a bigger size than the two previously published models. However, these forecasts also address the Senate, and they suggest Republicans are favored to retain control of the upper chamber. Those two models are described in this week’s Crystal Ball. But first, Campbell has an introductory piece explaining these models and the overall midterm environment. Explanations of all of these models, as well as Campbell’s introduction, will be appearing in PS: Political Science and Politics. We’d like to thank Campbell for once again convening this collection of forecasts, which is something he does for the Crystal Ball and PS every two years. For details about the four forecasts included in this year’s collection, click on the names above.

The Editors

On Tuesday, Nov. 6, about 90 million American voters (around 40% of the voting-eligible population, give or take) will elect all 435 members of the U.S. House of Representatives and 35 members of the U.S. Senate.[1] The midterm election’s outcome will play a major role in policy-making and the politics leading up to the presidential election of 2020. Going into the 2018 elections, Republicans hold majorities in both chambers of Congress. This collection of four different models printed in the Crystal Ball offers forecasts of how the 2018 midterm congressional elections are likely to change the partisan composition of the House and the Senate.

As in previous forecasting symposiums, the forecasts in this collection are based on statistical analyses of how indicators of the pre-campaign context of elections have been associated historically with their outcomes. While each model draws on its own set of indicators, some context may be helpful. What do the history of midterm elections and the number of seats each party is defending in 2018 (the arithmetic of the election), and the political climate leading up to the fall campaign broadly indicate about 2018?

Midterm history

As is well known, the president’s party routinely loses House seats in midterms. The presidential party has lost seats in all but three (1934, 1998, and 2002) of the 29 midterms held since 1900. In the 17 midterms held since 1950, presidential party losses averaged 24 House seats. Since the mid-1980s, the mean loss has been 20 seats and a median of only 10 seats.

While midterm losses in House seats are routine, they vary greatly, and this variation has increased substantially. The standard deviation of presidential party midterm seat losses increased from a standard deviation of 17 seats (1950 to 1982) to 27 seats (1986 to 2014).   Polarized parties have diminished the number of easily flipped seats; but, in doing so, they have increased the number of seats taken when short-term conditions more strongly favor one party. House seat changes are in an era of feast or famine. Based on the history of seat changes, the real question is not which party will gain seats, but whether Democratic House seat gains will be small or large.

The congressional arithmetic of 2018

The number of seats each party holds and needs to win to achieve a majority, the congressional arithmetic of 2018, is also an important context of the election. At first glance, current party divisions would seem to favor Democrats in the Senate and Republicans in the House. However, the prospects of Democratic majorities are just the opposite: greater in the House than in the Senate. Despite their presidential candidate losing the national popular vote and losing a net six seats, Republicans won a 24-seat majority (241 Republicans to 194 Democrats) in the House in 2016. So, using the prior election as the baseline, Democrats need a pickup of 24 seats to reach 218 seats for a majority.[2] As the midterm history reviewed above indicates, a midterm loss for Republicans of this magnitude would not be especially unusual.

In the Senate, Republicans hold a slim majority (51 to 49, counting two nominal “independents” who caucus with the Democrats). Taking into account Vice President Mike Pence’s tie-breaking power, Democrats require only a two-seat gain for majority control. This seemingly small shift, however, is a tall order. Democrats are defending many more seats than Republicans this year. Of the 35 Senate seats up, Democrats hold 26 and Republicans just nine. This does not offer much room on the “high side” for Democrats or much room on the “low side” for Republicans. Democrats have to do well just to hold their current overall numbers.

The reason for this is found in the previous three elections for this class of Senate seats. In this Senate class’s last election in 2012, Democrats registered a small net gain (two seats), but this was in addition to their strong showings in the prior two elections for the class. In 2006, Democrats gained six seats, and in 2000, they gained four seats. A net gain of two seats and a majority in the Senate requires means Democrats have to win a whopping 80% (28 of 35) of this year’s Senate elections.

The political climate of 2018

That is history, but how is the current political climate steering the electorate? Despite strong economic growth (real GDP growth of 4.2% in the second quarter), federal income tax cuts, and a low and declining unemployment rate (under 4%), the political climate leading into the midterm seems to favor the Democrats. An important indicator of this is the president’s approval rating. In mid-August, President Donald Trump’s approval hovered around 40%. Is this good or bad news for the in-party?

The good news for Republicans is that then-candidate Trump’s numbers had not been high in 2016 when voters last elected a Republican House and Senate. If Republicans in Congress could survive candidate Trump in 2016, why can’t they survive President Trump in 2018? Moreover, aside from Presidents George H.W. and George W. Bush, whose approval ratings had been inflated by Desert Storm and the 9/11 terrorist attacks, President Trump’s current numbers are not much different from the ratings of other recent presidents going into their first midterm. Other than the Bushes, every president since the mid-1970s has had approval ratings in the low to mid 40s at their first midterm. As of this week, Trump’s Gallup approval rating stands at 40%.

The bad news for Republicans is that the parties of each of these four recent (non-Bush) presidents (Jimmy Carter, Ronald Reagan, Bill Clinton, and Barack Obama) suffered double-digit House seat losses in their midterms and two lost more than 50 seats (Clinton in 1994 and Obama in 2010).

The forecasts

With the history, arithmetic, and climate considered, we can now turn to the systematic estimates of how these factors are likely to come together in this year’s election. What follows are four independent congressional forecasts that inform us about what we should expect to come out of this election. Table 1 presents a summary of these congressional forecasts. Although there are differences among them, two points are common. This is likely to be a very good year for the Democrats in the House of Representatives. In fact, all four forecasts expect a Democratic House majority. In the Senate, Republicans are likely to hold their own and perhaps pick up a seat or two. So, on to the forecasts. For more details about the forecasts, see the links at the top of this piece.

Table 1: Summary of the 2018 midterm congressional election forecasts

James E. Campbell is the UB Distinguished Professor of Political Science at the University at Buffalo. His most recent book is Polarized: Making Sense of a Divided America.


1. Mississippi’s special nonpartisan election on Nov. 6 for its Class II seat seems likely to result in a runoff election on Nov. 27.

2. After now-Rep. Conor Lamb (D, PA-18) won a March special election, Democrats need 23 net seats to win a majority in the U.S. House.

The Seats-in-Trouble Forecasts of the 2018 Midterm Congressional Elections

James E. Campbell, Guest Columnist September 13th, 2018


The Seats-in-Trouble model of party seat change in national congressional elections (both on-year and midterms) is a hybrid election forecasting model. It combines the insights and comprehensive assessments of expert election analysts examining in depth the conditions of individual House and Senate contests with a rigorous statistical analysis of historical aggregate data of partisan seat change.

From seats exposed to seats in trouble

The model essentially builds on the seat exposure idea explored by Bruce Oppenheimer, James Stimson, and Richard Waterman and similar notions examined by others who recognized the simple facts that you can’t lose what you don’t have and you are more likely to lose a lot if you have a lot to lose. Taking it to the next level, if you have a lot of vulnerable seats, chances are greater you’ll lose a lot of seats. The Seats-in-Trouble forecasting equation built from these observations was first used in the 2010 House elections. It has since been modified, tweaked, and extended to Senate elections. The accuracy of its forecasts in its various incarnations has been quite good until 2016, when substantial expected Democratic seat gains failed to materialize. As a result, the model’s standard of what constitutes a “seat in trouble” has now been tightened.

From individual ratings to an aggregate index

The raw material of seats-in-trouble equation is the pre-election competitiveness ratings of individual district or state races determined by The Cook Political Report since the mid-1980s.[1] Races are rated in one of eight categories: solid, likely, or leaning to the Democrats, a Democratic toss-up seat, and the same four levels of competitiveness on the Republican side. The seats-in-trouble index aggregates these individual district or state (in application to the Senate) ratings to a national measure. The revised index now counts a seat as being in trouble if it is rated as a toss-up or worse for the party currently holding the seat. The index is computed as the net difference in the number of seats in trouble for each party: Democratic seats in trouble minus Republican seats in trouble. A negative relationship is expected between the number of seats in trouble for a party and the number of seats actually gained by that party.

From an aggregate index to a statistical forecast

How has the net number of a party’s seats in trouble in August of an election year translated historically into the net number of seats it won or lost in November? The seats in trouble index is plotted against seat change for Democrats in House and Senate elections in Figures 1 and 2 respectively, and the associated regression equations are presented in Table 1.

Figure 1: Democratic seat change in the House by seats in trouble, 1984-2016

Note: Ratings around August were not made in 1986 and 1990 and those elections, therefore, are not included.

Figure 2: Democratic seat change in the Senate by seats in trouble, 1988-2016

Note: Ratings around August were not made in 1990 and are, therefore, not included.

Table 1: The Seats-in-Trouble forecasting equations of Democratic Party congressional seat change

Notes: N=15 for House, 14 for Senate. **p<.01, one-tailed. *p<.05, one-tailed. Standard errors are in parentheses. House estimates are based on elections from 1984, 1988, and from 1992 to 2016. Data was unavailable for the Senate in 1984. The seats in trouble count is from The Cook Political Report in or around mid-August of the election year. The 2018 reading was of The Cook Political Report on Aug. 17, 2018.

The House forecast

As Figure 1 and the table indicate, the index and actual seat changes are strongly related in House elections and the conversion of vulnerable seats to lost seats is greater than one in absolute value. If there are strong partisan short-term forces developing in an election, August ratings pick up their early signs, but not their full brunt as they build to November. In concrete terms, for every two “toss-up or worse” seats a party has in August compared to its opposition, it should expect to lose three seats in November.

The Cook Political Report in mid-August 2018 rated three House seats held by Democrats and 37 seats held by Republicans as toss-ups or worse. With a toss-up or worse difference of -34 for the Democrats, they are likely to gain about 44 seats. If this comes to fruition, the next House will have a Democratic majority with about 238 Democrats and 197 Republicans. Based on both in sample and out of sample errors, the odds are very high that the Democrats will emerge from the midterm with control of the House.

The Senate forecast

The association between Senate seats rated in trouble in August and actual seat change is also strong, though not quite the level of that in the House. The conversion of seats in trouble to Senate seat change is nearly one-to-one.

Because Democrats are defending a much larger number of Senate seats than Republicans this year, there are many more Democrats seats that could potentially be in trouble. As of mid-August, Cook rated five of the 26 Senate seats defended by Democrats and three of the nine Senate seats held by Republicans as toss-ups or worse. With Democrats having two more Senate seats in trouble than Republicans, we should expect Republicans to pick up two more seats, leaving the next Senate with 53 Republicans and 47 Democrats.

James E. Campbell is the UB Distinguished Professor of Political Science at the University at Buffalo. His most recent book is Polarized: Making Sense of a Divided America.


1. Thanks to Charlie Cook and his colleagues at The Cook Political Report for generously sharing their data.

Congressional Forecasts for 2018: Structure-X Models

Charles Tien and Michael S. Lewis-Beck, Guest Columnists September 13th, 2018


We build here on our “Structure-X model,” successfully applied in 2014. We first generate a 2018 forecast from our classic structural model. Next, we adjust this forecast on the basis of expert judgments provided in Inside Elections with Nathan L. Gonzales.

With respect to the first step, the structural model, we employ the following equation:

House Seat Change = Presidential Approvalt–1 + Disposable Incomet–1 + Mid-termt  (Eq. 1)

OLS yields the following results,

HS = -44.83 + 0.82*P + 4.91*I – 28.53*M  (Eq. 2)

R2 = .59, adj. R2 = .55, RMSE = 18.06, D-W = 1.87, N = 35

Notes: HS = presidential party seat change in the House of Representatives, I = change in real disposable income, for initial six months of the election year (from the Bureau of Economic Analysis’s National Income and Product Account Table 2.1: Personal Income and Its Disposition), P = June Gallup poll presidential popularity rating from Gallup’s Presidential Approval Center, M = midterm dummy (0 = presidential election, 1 = midterm election), figures in parentheses are t-scores, * = statistical significance beyond .05.

To forecast the 2018 seat change in the House, we insert the independent variable values: I = 1.73 (December 2017 through June 2018), P = 42 (June 2018), and M = 1:

HS2018 = -44.83 + 0.82 (42) + 4.91 (1.73) – 28.53 (1)  (Eq. 3)

= -30 Republican seats.

As a second step, we now modify that forecast by taking into account expert judgments, using the total “seats in play” number drawn from Inside Elections. For June 2018, Gonzales reports 68 Republican seats in play, versus 10 Democratic seats in play. Thus, the X-number = (10 – 68) = – 58.   This expert index (-58) clearly departs from our structural forecast (-30).  To reconcile these differences, we average the two methods, meaning a June-based forecast of (-30 – 58)/2 = -44 net Republican seat loss.

We also use the same Structure-X modeling strategy to forecast the 2018 Senate midterm election. First, we use our classic Senate structural model to estimate a midterm forecast:

Senate Seat Change = Presidential Approvalt–1 + Disposable Incomet–1 + Mid-termt + Seats Exposedt  (Eq. 4)

OLS yields the following:

SC = 2.54 + 0.14*P + 0.90*I – 2.43*M – 0.69*Expose  (Eq. 5)

R2 = .68, adj. R2 = .64, RMSE = 2.89, D-W = 1.85, N = 35

where SC = presidential party seat change in the U.S. Senate, Expose is number of seats the president’s party has up for reelection, and the other variables are as defined in Equation 2. Inserting the 2018 independent variable values produces the Senate forecast of:

SC2018 = 2.54* + 0.14*(42) + 0.90*(1.73) – 2.43*(1) – 0.69*(9)  (Eq. 6)

= 0 (no seat change)

We adjust this structural model forecast by adding Gonzales’ expert state-level evaluation of races. For July 2018, Gonzales estimates that there are four Democratic party seats in trouble and two Republican seats in trouble. A simple averaging of the structural forecast (zero) and expert judgement (two) produces a Structure-X forecast of a one-seat pickup for the Republicans in the Senate.

Our Structure-X model forecasts that the midterm elections will result in a divided government. The Democratic Party is forecast to win 44 seats. This outcome would give the Democrats a total of 239 House seats and control of the speakership and all House committees. The newly seated House of Representatives controlled by Democrats would be balanced by a Republican Senate and executive branch.

Charles Tien is Professor of Political Science at Hunter College and the Graduate Center, CUNY. Michael S. Lewis-Beck is F. Wendell Miller Distinguished Professor of Political Science at the University of Iowa and co-author of the book Forecasting Elections.