This is part one in a series on the 2017 France presidential election. Read part two here.
Election day in France is nearly upon us, but before I write an article about our final forecast for the race I want to cover why that model is important. First, read this post. Then, check out my polling average and election forecast. After that, wait for Part 2 to be posted on Saturday!
With that in mind, let’s start with a quick look at where the race for president stands.
The State of The Race
The current state of the race in France is, well, complicated. The current “leader” in the polls (to the extent that we can discern a leader — more on that below) is Emmanuel Macron. Macron is a former banker and past member of the Socialist Party in France. Now, however, he is the founder of his own party En Marche!, which is a progressive political party intended to “transcend” traditional partisan divides. Some people call this “trans-partisan,” but in the United States we call them independents (some people have called Macron a Michael Bloomberg type, but I don’t think that’s really a good comparison).
Like I mentioned, it’s hard to really call Macron the leader in the polls at this particular moment in time. On it’s face, polling shows that Macron has a roughly 1 percentage point lead over the far-right, anti-EU candidate Marine Le Pen. However, two days ago, she was in the lead. See the problem here? Polls in France are currently in a late-breaking flux, where we cannot directly (or even remotely definitively) identify a candidate with a clear lead. Nevertheless, Macron appears to have the highest number of supporters. So he’s the “leader.”
Two to three percentage points below the Macron-Le Pen entanglement is that of the traditionally Republican candidate François Fillon and far-left (“Bernie Sanders-of-France”) candidate Jean-Luc Mélenchon. Although Fillon has made some headlines for allegations that he created fake governments jobs to pay his family some big bucks, the rise of Jean-Luc Mélenchon is (in many ways) much more interesting.
For one thing, Mélenchon has surged in the polls, and I don’t use that term lightly. In fact, in comparative electoral context, that surge may be more of a tsunami. Jean-Luc Mélenchon has risen about 7% in our polling averagein the last 31 days. That’s absolutely incredible movement (although, not too uncommon for French elections) that calls into question whether “enthusiasm” of voters will play a big factor when le peuple go to the polls on Sunday. Again, I can’t underscore enough how absolutely bonkers this shift in public preferences is, especially in such a media-frenzied election. There’s also the consideration that politics is becoming more polarized, with people deciding hard-and-fast on their candidates earlier in time, which makes Mélenchon’s surge (ahem, “tsunami”) even more astounding.
In the end, though, these numbers alone don’t mean a lot to us; we are left with a lot of questions. Is Macron’s lead a stable lead? Is it even a safe lead? Is there a chance that Mélenchon is actually polling above the other three candidates? Could Fillon — mired in controversy — run away with Le Pen to a runoff wonderland?
Enter the forecast model.
Making Sense of the Numbers
As I mentioned above, a writeup on the final election forecast is slated to be published Saturday morning (no promises!) and so this section will cover the purpose and practical use of the forecast itself, but not (all of) what is says about the race for France president.
First, I want to include some of the “theory” behind the forecast model. I like to give credit where credit is due, and so I’ll include the two books by Nate Silver (of course) and “superforecaster” Joseph E. Tetlock from which most of these modeling underpinnings come. With that being said, let’s explain the usefulness of the forecast itself.
1. Outcomes are probabilistic.There is always some chance that an outcome can happen. Sometime that chance is zero, and sometimes it is one-hundred — but often times, the outcomes we feel most certain about have just an eighty percent chance of happening.
2. The output data is quantitative, but ought to be combined with qualitative analysis.Being told that a given presidency is only fifty percent likely is not very informative at face value. Instead of just spouting off numbers, analysts should utilize forecasts to explain what we can expect in certain events, often based on what (properly calibrated) forecasts have said in the past.
The forecast can also be useful in gauging the impact of certain events, or the wiggle-room candidates have when making important political choices. There are just a couple of the scenarios in which the probabilities of the forecast can be helpful.
3. The model responds to new data quickly, but keeps old numbers in mind.It only makes sense for us to take recent data more seriously than old — events can render old data obsolete — but it is often the case the political environments change without big events happening to spur that shift. Not only does this make sense, but doing so has helped our model make better predictions in the past.
This also causes the model to look rather volatile for some races. However, we would rather have this volatility than a model than treats information that is one week old the same as information coming out on the day before election day.
4. The past is a pretty good indicator of what we (don’t ) know about the future.The oft-repeated adage that “the past is not indicative of the future” may be right for some things: short term poker odds and votes on low-level congressional legislation, maybe. But in forecasting there is one thing the past conveys very well: error. We can use the average of past polling error as an indicator of the error, or uncertainty that we might see this year.
Finally, any good modeler will give you the following advice from statistician George Box: “all models are wrong, but some are useful.”
So what does this all add up to? Well, primarily, it’s a word of caution not to use the forecast as a hard and true expectation of the outcome of Sunday’s contest. The model ought not to be used as an absolute determiner of the election — that is to say, there is room for error in what the data tells us. Yes, Emmanuel Macron has roughly 23% of the vote in the France election. The above graphic shows that we can expect there to be error in the final-day election polls of 2.5%.
|Year||Poll Margin (%)||Election Margin (%)||Error (%)|
Further, when we take an average of those last-minute polls, error increase to 3% on average. That means that there has been polling errors both larger (and lower) than that 3%. In this case, an error of 3% would be enough to sink Macron, or it could be enough to ensure his procession to the second round. And again, as Mélenchon has surged in recent weeks, error could certainly be greater enough to put him in the head-to-head round.
What We’re Left With
In the end, analysts face a difficult choice with the France election. On the one hand, Emmanuel Macron is in a “superior” (again, just by 1%) position in the polls and has a good electoral position; usually, the candidate with the ability to appeal to the most voters has the most voters. In this case, centrist Macron may be that candidate.
On the other hand, polling has been wrong (twice) in the past and has certainly erred. When we combine the expectation of error within the bounds of what we’ve seen in the past, Macron’s “lead” becomes much less clear. Any number of possible match-ups (Macron v Le Pen, Fillon v Le Pen, etc.) are possible at this point in time. Of course, Macron versus Le Pen is the most likely outcome of the first round of the election — as they are the two leaders in the polls — but we can really only expect that outcome to happen a certain amount of the time.
The question is, how often?
The answer to that question and more in France 2017 Election Part Deux. Stay tuned, and in the meantime follow along on twitter