French presidential election: Experts diverge from polls with respect to Valérie Pécresse’s chances against President Macron

PollyVote has completed its January expert survey for this year’s French presidential election. With respect to the first round results, the expert consensus is pretty much in line with recent polls in that President Macron is likely to gain most of the votes. When it comes to the question of who will come in second, both forecasting methods expect a tight race between Marine Le Pen and Valérie Pécresse. While polls have Le Pen and Pécresse virtually tied, the combined expert forecast predicts a slight advantage for Pécresse.


With respect to the second round, PollyVote asked the expert panel to predict the outcomes of both hypothetical matchups. For a potential Macron-Le Pen matchup, experts and polls are in line and predict a rather comfortable 12-point margin for President Macron.


In case Macron would face Pécresse, however, polls and expert consensus diverge. While the polls also expect a Macron win (54% vs. 46%), the current expert consensus expects a very close race, with a slight advantage for Valérie Pécresse.


Once a month, the experts forecast the vote shares of the first round in this year's French presidential election, as well as the outcomes of hypothetical matchups. Five experts participated in the second round of the survey, which was conducted from January 18-21. The PollyVote team would like to thank Bruno Cautres (Sciences Po), Jean-Sébastien Ferjou (Atlantico), Mathieu Gallard (Ipsos), Bruno Jeanbart (Opinion Way), and Nicolas Prissette (Fondamental).

Forecasting panels at the APSA Annual Meeting

At this year’s APSA Annual Meeting, the Political Forecasting Group hosts four panels on election forecasting, all of which will be virtual:

Day

Time (PDT)

Session title

Discussant

Speakers

Wed, 9/29

6:00 –
7:30 am

Political Forecasting: Teaching and Practice

Ross Burkhart

  • Lara Brown, Lilly Goren &
    Edward Levitas
  • Debra Leiter
  • Philippe Mongrain

Thu,
9/30

10:00 -11:30 am

Lessons learned from the 2020 U.S. presidential election forecasts (Roundtable)

Mary Stegmaier (Moderator)

  • Ruth Dassonneville
  • Charles Tien
  • Andreas Murr
  • Veronique Jerome
  • Thomas Rietz

Fri,
10/1

2:00 –
3:30 pm

Political Forecasting Methodology

Peter Enns

  • Michael Lewis-Beck & Charles Tien
  • Carl Klarner
  • Ross Burkhart

Sun,
10/3

6:00 –
7:30 am

Forecasting the 2021 German federal election

Debra Leiter

  • Bruno Jerome, Veronique Jerome
    & Michael Lewis-Beck
  • Mary Stegmaier
  • Andreas Murr
  • Andreas Graefe

Don’t miss this. In addition, you may want to check out another session on forecasting elections not hosted by the Political Forecasting Group, which will take place in-person on Thursday, September 30 from 12:00-1:30pm.

Final expert survey: Experts predict a dead heat between SPD and CSU/CSU in run for strongest party

PollyVote has completed the fifth and final round of its monthly expert survey for this year’s German federal election. The experts expect a dead heat, with SPD gaining 24.2% of the vote, slightly ahead of the CDU/CSU, which is predicted to gain 23.6% of the vote. The Greens are expected to come in third (16.7%), followed by FDP (11.2%) and AfD (10.8%), who are virtually tied for fourth place. The Left (6.6%) is also expected to pass the 5%-electoral threshold. The predicted vote share for all other parties combined is 6.8 points.

Once a month, the experts forecast the vote shares of the respective parties in the federal election on September 26. A total of 19 experts participated in the fifth round of the survey, which was conducted from September 21-24. The PollyVote team would like to thank:

  1. Prof. Dr. Kai Arzheimer
  2. Prof. Dr. Joachim Behnke
  3. Dr. Evelyn Bytzek
  4. Prof. Dr. Marc Debus
  5. Prof. Dr. Martin Elff
  6. Prof. Dr. Simon Tobias Franzmann
  7. Prof. Dr. Oscar Gabriel
  8. Prof. Dr. Achim Goerres
  9. Prof. Dr. Max Kaase
  10. Dr. Michael Meffert
  11. Prof. Dr. Dieter Roth
  12. Prof. Dr. Peter Schrott
  13. Dr. Martin Steinbrecher
  14. Prof. Dr. Andreas Wüst
  15. Prof. Dr. Thomas Zittel

as well as four experts who preferred to remain anonymous.

Expert survey: SPD gains more than three points at the cost of CDU/CSU

PollyVote has completed the fourth round of its monthly expert survey for this year’s German federal election. The experts still expect the CDU/CSU to be the strongest party with 26.6% of the vote. However, the party lost 3.4 percentage points compared to the last survey round in July. For the first time, the expert forecast now has the SPD in second place with 19.9% of the vote (a gain of 3.3 points), slightly ahead of the Greens (19.6%, down from 20.4%). The FDP are at 11.1%, the AfD at 10.5%, and the Left Party at 6.8%. The predicted vote share for all other parties combined is 5.8 points.

 

Once a month, the experts forecast the vote shares of the respective parties in the federal election on September 26. A total of 19 experts participated in the fourth round of the survey, which was conducted from August 16-19. The PollyVote team would like to thank:

  1. Prof. Dr. Joachim Behnke
  2. Dr. Evelyn Bytzek
  3. Prof. Dr. Marc Debus
  4. Prof. Dr. Martin Elff
  5. Prof. Dr. Simon Tobias Franzmann
  6. Prof. Dr. Oscar Gabriel
  7. Prof. Dr. Achim Goerres
  8. Dr. Alexander Jedinger
  9. Prof. Dr. Max Kaase
  10. Dr. Michael Meffert
  11. Prof. Dr. Dieter Roth
  12. Prof. Dr. Peter Schrott
  13. Dr. Martin Steinbrecher
  14. Prof. Dr. Andreas Wüst
  15. Prof. Dr. Thomas Zittel

as well as four experts who preferred to remain anonymous.

New book: Laws of Politics

Alfred G. Cuzán (University of West Florida), co-founder of the PollyVote (from which he retired last year), has a book named Laws of Politics, scheduled for publication in August. Here is the description from the publisher’s website:

Drawing on classic and contemporary scholarship and empirical analysis of elections and public expenditures in 80 countries, the author argues for the existence of primary and secondary laws of politics.

Starting with how basic elements of politics—leadership, organization, ideology, resources, and force—coalesce in the formation of states, he proceeds to examine the operations of those laws in democracies and dictatorships. Primary laws constrain the support that incumbents draw from the electorate, limiting their time in office. They operate unimpeded in democracies. Secondary laws describe the general tendency of the state to expand vis-à-vis economy and society. They exert their greatest force in one-party states imbued with a totalitarian ideology. The author establishes the primary laws in a rigorous analysis of 1,100 parliamentary and presidential elections in 80 countries, plus another 1,000 U.S. gubernatorial elections. Evidence for the secondary laws is drawn from public expenditure data series, with findings presented in easily grasped tables and graphs. Having established these laws quantitatively, the author uses Cuba as a case study, adding qualitative analysis and a practical application to propose a constitutional framework for a future Cuban democracy.

Written in an engaging, jargon-free style, this enlightening book will be of great interest to students and scholars in political science, especially those specializing in comparative politics, as well as opinion leaders and engaged citizens.

Expert survey: CDU/CSU gains at the cost of the Greens

PollyVote has completed the third round of its monthly expert survey for this year’s German federal election. The experts expect the CDU/CSU to be the strongest party with 30.0% of the vote, a gain of nearly three points compared to the last survey round. The Greens lose 2.4 points and are now predicted to receive 20.4%. The SPD is expected to achieve 16.6% (+0.3 points), followed by the FDP at 10.5%, the AfD at 9.9%, and the Left Party (7.2%). The predicted vote share for all other parties combined is 5.4 points.

Once a month, the experts forecast the vote shares of the respective parties in the federal election on September 26. A total of 17 experts participated in the third round of the survey, which was conducted from July 7-12. The PollyVote team would like to thank:

  1. Prof. Dr. Kai Arzheimer
  2. Prof. Dr. Joachim Behnke
  3. Dr. Evelyn Bytzek
  4. Prof. Dr. Marc Debus
  5. Prof. Dr. Simon Tobias Franzmann
  6. Prof. Dr. Oscar Gabriel
  7. Prof. Dr. Achim Goerres
  8. Dr. Alexander Jedinger
  9. Prof. Dr. Max Kaase
  10. Dr. Michael Meffert
  11. Prof. Dr. Peter Schrott
  12. Prof. Dr. Andreas Wüst
  13. Prof. Dr. Thomas Zittel

as well as four experts who preferred to remain anonymous.

PollyVote expert survey: CDU/CSU five points ahead of the Greens

PollyVote has completed the second round of its monthly expert survey for this year’s German federal election. The experts expect the CDU/CSU to be the strongest party with 27.4%, roughly 5 points ahead of the Greens, which the experts see at 22.7%. The SPD is expected to gain 16.3%, followed by the AfD and FDP (10.6% each), and the Left Party (7.0%). The predicted vote share for all other parties combined is 5.5 points.

Once a month, the experts forecast the vote shares of the respective parties in the federal election on September 26. A total of 17 experts participated in the second round of the survey, which was conducted from May 28 to June 1. The PollyVote team would like to thank:

  1. Prof. Dr. Kai Arzheimer
  2. Prof. Dr. Joachim Behnke
  3. Dr. Evelyn Bytzek
  4. Prof. Dr. Marc Debus
  5. Prof. Dr. Martin Elff
  6. Prof. Dr. Simon Tobias Franzmann
  7. Prof. Dr. Oscar Gabriel
  8. Prof. Dr. Achim Goerres
  9. Dr. Alexander Jedinger
  10. Prof. Dr. Max Kaase
  11. Dr. Michael Meffert
  12. Prof. Dr. Dieter Roth
  13. Prof. Dr. Peter Schrott
  14. Dr. Markus Steinbrecher
  15. Prof. Dr. Andreas Wüst

as well as two experts who preferred to remain anonymous.

First expert survey for German federal election: Greens six point behind CDU/CSU

PollyVote has completed the first round of its monthly expert survey for this year’s German federal election. The experts expect the CDU/CSU to be the strongest party with 28.7%, nearly 6 points ahead of the Greens, which the experts see at 23.0%. The SPD is expected to gain 16.7%, followed by the AfD (9.8%), the FDP (9.3%), and the Left Party (7.8%). The predicted vote share for all other parties combined is 4.7 points.

Once a month, the experts forecast the vote shares of the respective parties in the federal election on September 26. A total of 19 experts participated in the first round of the survey, which was conducted from April 27-30. The PollyVote team would like to thank:

  1. Prof. Dr. Kai Arzheimer
  2. Prof. Dr. Joachim Behnke
  3. Dr. Evelyn Bytzek
  4. Prof. Dr. Marc Debus
  5. Prof. Dr. Martin Elff
  6. Prof. Dr. Simon Tobias Franzmann
  7. Dr. Heiko Giebler
  8. Prof. Dr. Achim Goerres
  9. Dr. Alexander Jedinger
  10. Prof. Dr. Max Kaase
  11. Dr. Michael Meffert
  12. Prof. Dr. Dieter Roth
  13. Prof. Dr. Peter Schrott
  14. Dr. Markus Steinbrecher
  15. Prof. Dr. Andreas Wüst
  16. Prof. Dr. Thomas Zittel

as well as three experts who preferred to remain anonymous.

Call for Paper proposals: PS: Political Science & Politics Symposium on ‘Forecasting the 2021 German Federal Election’

With the support of the APSA Political Forecasting Group, Bruno Jérôme and I are putting together a symposium of papers on “Forecasting the 2021 German Elections”, which will be published in PS: Political Science & Politics. We are interested in a diversity of approaches to forecasting (e.g., poll aggregation, betting markets, structural models, or combinations thereof), including ways to communicate the uncertainty of forecasts. The symposium will have space for updates of models that have been used historically to predict German elections, but we also encourage submissions that lay out new and innovative ways to generate and communicate election forecasts and their underlying uncertainty. We also are keen to include voices that might be critical of election forecasting. 

We are seeking short papers of 3,000 words maximum, and particularly welcome submissions from diverse teams of scholars and members of underrepresented groups as well as junior scholars. Membership in the APSA Political Forecasting Group is by no means a requirement. Abstracts should be submitted by 31 March 2021 to the guest editors Bruno Jérôme (bruno.jerome@gmail.com) and Andreas Graefe (graefe.andreas@gmail.com), who will evaluate those abstracts and will make a pre-selection of papers for possible inclusion in the Symposium proposal. If your proposal is accepted, submissions of full papers are due some time in May (exact date still to be determined), and final decisions about the acceptance of papers will be made after the review process. To ensure a timely publication of the Symposium, authors have to commit to responding to reviews in one-week time. For questions regarding the Symposium proposal, contact Bruno and myself.

Post-election analysis: Accuracy of popular vote forecasts

As in previous elections since 2004, PollyVote once again did a very good job in predicting the popular vote. In its final forecast, the PollyVote predicted that Joe Biden will gain 52.2% of the two-party popular vote, a forecast that has remained remarkably stable since it first publication on May 15. Although votes are still being counted, and we probably won’t know the actual result for several weeks, Biden’s two-party vote will likely end up around 52% – very close to what PollyVote predicted.

PollyVote did well compared to the closely watched forecasts published by FiveThirtyEight and the Economist. Due to their heavy reliance on polls, which substantially overestimated Biden’s lead not only on the state level but also nationwide, these models predicted that Biden would win the popular vote by eight to nine points. The following chart, which will automatically update until the final vote count is known, shows the extent to which the PollyVote outperformed these benchmarks over the course of the campaign.

 

It’s only one election. Did PollyVote just get lucky?

PollyVote’s 2020 performance is well in line with previous elections, as we have shown in our research, which has been awarded with an “Outstanding Paper Award” by the International Journal of Forecasting.

The average error of the PollyVote’s popular vote forecast across the last 100 days prior to the seven elections from 1992 to 2020 (ex post for the three elections from 1992 to 2000) was only 1.1 points, and more accurate than forecasts from any other method.

Now, let’s take a look at the historical performance of the PollyVote in predicting the popular vote, compared to well-known benchmarks methods that also have a few elections on their belts.

The following chart compares the accuracy of the PollyVote and FiveThirtyEight across the three elections from 2012 to 2020 for which we have data from both forecasts (FiveThirtyEight was first launched in 2008 but we don’t have historical data available). For each single day, the chart shows the average absolute error one would have achieved when relying on the PollyVote or the FiveThirtyEight model from that day until Election Day – across (and in each) of the three elections from 2012 to 2020. The line for the PollyVote is consistently below the respective line for FiveThirtyEight, which means that, on average, the PollyVote forecasts were more accurate. For example, across the last four weeks (or 28 days) across the last three elections, the PollyVote’s average error was 1.2 percentage points, compared to 1.6 points for FiveThirtyEight.

Now, let’s compare the PollyVote’s forecast accuracy to a simple polling average. Here, we use RealClearPolitics, which allows us to compare forecasts back to the 2004 election, which is when the PollyVote was first launched. Again, on average, one would have clearly fared better by relying on the PollyVote instead of the RCP poll average. For example, across the last four weeks (or 28 days) across the last five elections, the PollyVote’s average error was 0.9 percentage points, compared to 1.5 points for the RCP poll average.

 

These accuracy gains of roughly half a point may seem small in absolute terms. However, given the closeness of recent elections, errors of that magnitude can make a big difference. Also, if one looks at the percentage of error reduction, the improvement in forecast accuracy is substantial, as can be seen in the chart on the right.

For example, if one had relied on the PollyVote instead of the FiveThirtyEight model for the last four weeks before the elections from 2012 to 2020, one would have reduced the forecast error by roughly 30%. 

Compared to the RealClearPolitics poll average across the five elections from 2004 to 2020, error reductions are even higher and exceed 40%, depending on the time remaining to Election Day.

Benefits of combining forecasts

The PollyVote relies on a fundamental principles from nearly half a century of forecasting research: combining forecasts that rely on different methods and different information reduces error. This approach has two major advantages.

Combining protects from picking a poor forecast

The relative accuracy of different forecasting methods often varies across time. That is, methods that have worked well in one election might not work well in another, and vice versa. For example, polls performed pretty well in 2012, OK (at least at the national level) in 2016, and poorly in 2020. When predicting a single event (e.g., the 2020 election), a combined forecast such as the PollyVote will always be at least as accurate as the typical component forecast. Therefore, the PollyVote prevents the forecaster from making large errors when relying on a particular method that ends up being far off, such as polls in 2020.

Combining likely to prevail in the long run

Combining forecasts reduces error by canceling out systematic and random errors of individual forecasts, an effect that is particularly strong if the individual forecasts rely on different methods and data. If the different component forecasts “bracket” the final outcome, the combined forecast will be among the most accurate forecasts available. In the long run (e.g., when being used across many elections), it is difficult to think of a better way to forecasting than by combining different methods that use different information.