German Election Update Stay the Course
The German Election is in less than two months. Traditional forecasts, such as polls and betting markets, have been fairly stable over the last month and continue to point to the CDU/CSU winning the election and the SPD losing ground in comparison to its current position. Social Media Influence (SMI) forecasts confirm polls and betting markets regarding the CDU/CSU and SPD but also continue to highlight expected gains for the FDP. Political analysts appear to be coming to a similar conclusion, namely that the CDU/CSU will win the largest portion of the vote and end up forming a new ruling coalition with the FDP pushing the SPD out.
Over the past month, expectations appear to have solidified. However, there is still time for shifts as in Germany election campaigning only starts 6-weeks before election-day, so campaigning starts on August 13 with election-day on September 24. During this intense 6-week period, which includes a televised debate on September 3, parties will aggressively get their messages out which leaves open the possibility of volatility.
Polls
Polls have remained fairly consistent during June and July. The leading party, the CDU/CSU, has maintained a range from around 37% to 40% whereas second place SPD has hovered around 23% to 25%. The four minor parties, AfD (populists), FDP (classical liberalism), Green (greens), and Left (far-left) all average in the high single digits to very low double digits and tend to vary little. From the perspective of polls, there have been few real changes over the last few months. These could start to shift once campaigning officially starts on August 13, but as for now poll figures appear stable.
Betting Markets
Betting markets by their nature tend to be more volatile than polls as these live markets can incorporate shifts of perception and news events instantaneously. However, even betting markets have remained stable over the last two months. For instance, the probability of Merkel being reelected German Chancellor hit a low of 84% and a high of 89%, according to Predictit during June and July. This range is extremely tight considering we are fairly close to the election and there has been plenty of relevant news that could have impacted betting markets.
Traditional Political Analysis
The current ruling coalition includes the two largest parties, the center-right CDU/CSU and the center-left SPD. Though not overly problematic, this coalition tended to prevent major decisions on a number of topics as the SPD offset moves by the CDU/CSU. In other words, as the coalition included the largest right-leaning and left-leaning parties, Germany ended up without a strong inclination to move in either direction.
Analysts have been calling for the next ruling coalition to include the CDU/CSU and perhaps one or two smaller parties. This appears feasible as CDU/CSU alone is polling just around 40% and the smaller parties are polling around 10%. In order for the next ruling coalition to include two parties, the CDU/CSU would have to perform at the upper end of its range and so would one of the smaller parties. Another likely outcome would be for the CDU/CSU to partner with two of the smaller parties.
The current front-runner or preferred coalition partner is seen as the FDP. Historically, this party has been the king-maker in Germany as it has acted on many occasions since the end of WWII as the coalition partner for either CDU/CSU or SPD. As a classical liberal party, the FDP has elements of a right-leaning party (pro-business and fiscally conservative) and left-leaning (socially liberal) so has been able to effectively partner with various parties.
The possibility of the FDP being the preferred or likely coalition partner for the CDU/CSU really began to take-off in June . Much of the logic centers on the compatibility of two parties in terms of policy. However, another reason for the FDP being highlighted and not one of the other small parties appears to be the increasing popularity of its leader, Lindner.
Social Media Influence (SMI)
As opposed to polls that measure a small randomly selected subset of the electorate and betting markets that allow speculators who follow politics and related news items to bet on candidates, SMI focuses on social media which allows for a much greater sample size of people in a less biased environment. The advantages of SMI over other forms of quantitative analysis will not be fully covered here, but it is important to highlight that SMI has tended to outperform during elections where there is either a strong bias for/against certain candidates in the media or there are strong shifts in the electorate’s preferences that go outside preconceived or past patterns. As elections have tended to fall into these categories over the last few years, SMI has tended to make accurate, though often out-of-consensus, forecasts.
Currently, SMI in Germany is heavily pointing towards the FDP as the party that will exceed current expectations. The leading CDU/CSU should perform more or less in-line with polls. And, other parties, including the SPD, should lose ground going into the election.
Chart1: German Election Polls versus SMI
Sources: ZettaCap, Infratest dimap, YouGov, INSA, Forsa, Forschungsgruppe Wahlen Note: the poll average was calculated using the simple average of the last five available polls as listed here.
The left portion of the chart shows that according to polls there are two principal parties, the CDU/CSU and the SPD, and four minor parties that clump together. As previously stated, these polls have been fairly stable for the last few months.
SMI ratings, on the right portion of the chart, show an alternate picture of the race as the SPD is pushed into a distant third place and the FDP is moved to a strong second. Put another way, the FDP has a tremendous amount of influence on social media, especially in comparison to its relatively weak poll figures. In other elections, such as in the US and France, such disparities between SMI and poll figures ended up with polls following SMI. So, in the case of Germany, FDP is expected to outperform its current poll figures to a significant degree. Such outperformance is not automatic or immediate, but requires time and enough exposure to that party or candidate.
FDP only party forecast to significantly outperform polls
The FDP has proven itself extremely strong in SMI terms, especially in relation to its poll figures. Comparing the two shows just how unusually strong the FDP is and, conversely, just how weak some of the other parties are.
Chart 2: SMI vs Average Poll Figures for main German Political Parties, as presented on X-axis from most left-leaning party on the farthest left of the chart to most right-leaning on the right
Source: ZettaCap, Infratest dimap, YouGov, INSA, Forsa, Forschungsgruppe Wahlen
In terms of expected outperformance, only the FDP stands out. The CDU/CSU comes extremely close to breaking even in that its SMI forecasts about the same level as polls. All other parties show significant expected underperformance. In theory, once the actual campaigning starts, we should see FDP poll figures increase going into the election.
An interesting side note is that the AfD, the nationalist party, posts fairly poor SMI. This argues against the idea that nationalist support is stronger than polls show due to the assumption that people are somehow less willing to publicly show support. In some elections, this phenomenon has occurred and often takes place around candidates who are perceived as being less socially acceptable. However, in the case of Germany and AfD, this does not appear to be the case.
Another interesting observation is that the Green Party is also posting very poor SMI. This is somewhat surprising for two reasons. First, the demographic make-up of the Greens in general tends to lend itself to social media in that they tend to be younger than the electorate. However, even with this built-in advantage, its SMI still struggles. Second, Austria, a neighbor of Germany and very close culturally, elected its first green president in 2016. So, there are strong indications pointing to the Greens as producing relatively high SMI, but in reality its SMI has struggled. Using just SMI, the Greens could actually fall below the 5% minimum threshold in the election, knocking them out of the Bundestag.
SMI has not been stable
As opposed to polls and betting markets, SMI has fluctuated over the last few months. Most notably the CDU/CSU has surged ahead of FDP in terms of absolute SMI to take the lead, with FDP retaining a relatively strong SMI against other parties and the highest SMI in relation to its poll figures.
The SMI movement helps to describe much of what has occurred over the last few months. FDP was a forgotten party earlier this year and most analysts did not expect it to excel during the general election. SMI called for it to gain traction and it did, especially in terms of expectations. In June and July, analysts began to speculate that FDP would be strong enough to partner in the ruling coalition with the CDU/CSU. This was in stark contrast to the FDP’s image early in the year as an also-ran which did not even break the 5% minimum in the last general election.
FDP further gained attention as its chairman, Lindner, continued to grab headlines. Some articles pointed to similarities with the recently elected French president, Macron, while others highlighted his ability to seemingly pull the party back from the grave. So, in general, the added traction of the FDP over the last few months was pre-dated by its superior SMI.
CDU/CSU surged past the FDP in terms of SMI during the period of Germany hosting the G20 summit in Hamburg. The main party in the current coalition gained in SMI as Merkel was seen as competently handling what some called street riots. Additionally, Merkel’s performance on the world stage as well as her stance concerning Germany’s and Europe’s need to assert its independence allowed her and her party to gain on social media.
Interestingly, only FDP and CDU/CSU have really been able to gain any significant traction on SMI and these two parties have been the most heavily cited as entering the next ruling coalition. The other parties have generally been seeing their SMIs slide slowly.
Is there enough time for the FDP to catch-up?
With less than two months to election-day, there appears to be limited time for the FDP to improve its poll figures to be in-line with its SMI. Having said this, we have seen spikes in polls in the last two months before election-day. Most recently, we can point to Melenchon in France’s election where his poll figures spiked right before the election. In this case, Melenchon’s potential was identified well before his last-minute spike as his SMI was so much more attractive than his poll figures. Given the proper opportunity, in Melenchon’s case it was televised debates, his poll figures jumped higher to be more in-line with his SMI.
FDP could replicate a similar outcome to that of Melenchon. It has an opportunity during the last 6-weeks before election-day when campaigning is officially open. As FDP has been able to gain significant influence on social media, it should hold that they will perform better in polls as they are given more exposure to the broader electorate during the last 6-week push. There are limits, some of which are analyzed later in this post, to how much we should expect this disparity between SMI and poll figures to close. Regardless, given strength from the FDP, a ruling coalition between the CDU/CSU and FDP should be more easily reached.
Why SMI could be wrong in forecasting the German election
The SMI has done an exceptional job forecasting previous elections. It is the only known quantitative system to have correctly predicted Trump winning with the right number of electoral votes and the first to forecast Macron would win the French election. It would seem like such a system would not be too keen to provide disclaimers.
However, the German electoral system is significantly different than those of the US and France. In those countries, the electorate votes directly for the presidential candidate they want. In the case of the US, the candidate with the most votes wins and in France if a candidate does not exceed 50% then another round is held. In both of these cases, voters directly chose the most high-profile political official.
In Germany, the system differs in two main ways:
First, the electorate has two votes on election-day as they chose (1) a local representative to serve in the Bundestag (Germany’s national parliament, similar to the House of Representatives in the US), and (2) a party that they prefer to represent them. Half of the seats in the Bundestag go to those who were directly elected as the local representative and the other half of the seats is split proportionately based on the party votes gained, with some exceptions.
Second, the German Chancellor is elected by the Bundestag and not directly elected by the electorate. So, the first decision that the new Bundestag makes is normally to elect the Chancellor. By tradition, the Chancellor is, but not always, the chairperson of the party winning the most votes in the general election.
The German system makes elections very local and party based and (seemingly) was created to prevent a single charismatic leader from getting elected as Chancellor. In short, it makes it very ineffective for an individual to go around the established party system and its hierarchies and make a case directly to the electorate.
You can clearly see that the German system varies significantly from those of the US and France. Trump (US) and Macron (France) appear to have produced such large upsets in their respective elections not only because they were charismatic leaders but because of the electoral system. Both essentially went around established party procedures and courtesies and took their messages directly to the electorate. The SMI picked up on these trends extremely early and accurately as the SMI takes into consideration actions on social media and not a political party’s real-world network, membership, size and influence.
If the US had a system similar to that of Germany, Trump would have had to lobby representatives in the House to vote for him. Even more difficult to imagine, he would by tradition have to be the chairperson of the party that won the most votes. This would have required Trump to spend a considerable amount of time working his way up to Chairperson of the Republican Party. In such a scenario where the German electoral system was transplanted onto the US, someone like Jeb Bush, who performed very poorly in SMI and direct voting terms but very well in terms of endorsements from establishment politicians, would have been a stronger candidate than Trump. A similar case would have played out in France as Macron would have been knocked out by establishment politicians within the larger parties.
In the case of the current German election, the FDP’s star chairman, Lindner, fits a similar description in that he is charismatic and tends to do well when speaking directly to the electorate. In fact, his speeches appear similar to rallies of Trump and Macron. He tends to produce a certain level of energy with his excitement over proposing reforms and making political change that seems to lack when more establishment based politicians speak.
Lindner’s SMI also tends to resemble those of Trump and Macron. In each case, these previously lesser known politicians became very influential on social media very quickly, beating out seasoned politicians in absolute terms within a short period of time. The problem is that Germans will not vote directly for Lindner for Chancellor. Lindner needs to focus on selling his party and his party’s local representatives to the electorate. Taking his energy and message directly to the electorate will help, but even an outstanding SMI will not necessarily win his party seats. He has to transform his individual SMI into party votes.
Imagine the case of Trump campaigning in the same style as he did during the 2016 campaign but instead of asking for people to vote for him, he had to get them to vote for a local representative instead. The end result would be the same, as the local representative would then (in theory) vote him in at a later time – but, it would be very questionable if Trump, who relied heavily on direct appeal to the electorate, could have won if the vote was indirect. In such an indirect vote, it is doubtful there would be enough sustained enthusiasm to elect a non-traditional candidate.
This is the situation we have in Germany where Lindner is producing a lot of traction but as the vote is indirect, the chances of him pulling off a Trump or Macron style upset seems more remote. The SMI, as explained elsewhere, has been slightly modified to place a larger focus on political parties to better reflect the reality in Germany. However, no adjustments were made to incorporate the indirect nature of the vote. Given the significant degree to which Germany’s electoral system varies by diminishing the focus on national-level politicians, a stronger disclaimer is needed. This disclaimer should be taken more seriously as much of the reason for the FDP’s strong bounce-back this election is due to its chairman, Lindner.
Conclusions
The general conclusions of the changes perceived over the last month and of the general direction of the German Election include:
· Polls have been stable, showing that the CDU/CSU should win the election with relative ease,
· Polls show that the CDU/CSU strength is such that they should be able to form a ruling coalition with one, or perhaps two, smaller parties,
· Polls show the relative weakness of the traditionally strong SPD,
· Betting markets have been stable, showing that the chances of Merkel being reelected exceed 80%,
· SMI ratings have fluctuated more than polls or betting markets,
· SMI ratings for the CDU/CSU have increased the most in the last few months,
· SMI ratings forecast CDU/CSU winning as do polls and betting markets,
· SMI ratings forecast extremely and unexpectedly strong performance by FDP, implying FDP will catapult from a distant fourth place in current polls to a strong second place by election-day,
· There are disclaimers to this forecast, namely that the German system varies significantly from those of other countries which in theory increases the importance of establishment parties and politicians and lowers the importance of social media, especially in identifying outliers, with such a disclaimer FDP should still come in a strong third place and outperform its current polls which would open the door for it to enter the ruling coalition which is a significant upset considering it did not even break the 5% minimum in the previous general election,
· Regardless, SMI forecasts a strong showing for FDP and for the eventual ruling coalition to include CDU/CSU and FDP, while leaving the door open for SMI to over-estimate the upside of FDP,
· Political campaigning only starts on August 13, or 6-weeks prior to election-day, which allows for considerable volatility going forward.