Trump, Brexit, Renzi – The result of evil big data geniuses or just Confirmation Bias at work ?
This is a somewhat “off-topic” post but at the end there will be a turn to investing, I promise !!!
A Swiss magazine published an article on the Weekend, where a “Data scientist” claimed that a method that he developed had made Donald Trump President.
This method, which claims to identify (and of course influence) people by 5 basic character traits was used by a company called “Cambridge Analytica” in order to secure Trump’s victory. And by the way, they were responsible for Brexit too. The article seems to be quite popular, I received the link 5 times over the weekend from very different sources
There are a couple of Youtube videos about Cambridge Analytica, for instance here. But there is a longer one which is more interesting
In the video from September, they are claiming that they managed to make Senator Cruz popular from scratch. They also claim that they can segment down to very small groups and deliver them the right message in a way that will more or less “guarantee” results. As an example they showed how they targeted very small sub groups in Iowa with targeted ads about guns to make them vote for Cruz in the caucus.
The end of the video is quite spooky when the CEO says that they are supporting one of the presidential candidates and they are “looking forward what will happen in the next 7 weeks”.
The big question: Did they actually do it ?
So the big questions is: Did they actually “rig” the election ? Honestly, I have no idea but there is of course the possibility that they had an influence. However in my opinion there is little actual proof that it worked they way they claim but it surely is a hell of a marketing video.
However the far bigger problem with regard to this topic is the following:
Our society is still at a very early stage in learning to cope with “Online Social Media”.
For a long time, the mainstream media more or less determined what the “average person” might read or see or hear about anything. There was always the “peer-to-peer” contact but this was usually limited to a very small (local) area or group of people.
Especially if you had a more radical opinion about something, it was much harder to find someone who would confirm your view of the world. The main stream media would only give you the main stream. So in the old times, someone with a very strange opinion pretty soon became a loner when he would insist on his point of view.
With the internet and social media however, one can suddenly get in contact with people who have the same opinion who otherwise one had never encountered. In order to gain conviction about something, confirmation is important.
This leads us to the so-called “Confirmation Bias” which according to Wikipedia is defined as follows:
Confirmation bias, also called confirmatory bias or myside bias,[Note 1] is the tendency to search for, interpret, favor, and recall information in a way that confirms one’s preexisting beliefs or hypotheses, while giving disproportionately less consideration to alternative possibilities. It is a type of cognitive bias and a systematic error of inductive reasoning. People display this bias when they gather or remember information selectively, or when they interpret it in a biased way. The effect is stronger for emotionally charged issues and for deeply entrenched beliefs. People also tend to interpret ambiguous evidence as supporting their existing position.
In my opinion, the internet and especially Social Media like Facebook are a huge “Confirmation Bias Machines”. Facebook will send you news based on what kind of news you clicked before and Google will show you those search results that resemble closest what you have searched before. I think there might be even a kind of “vicious circle” at work that motivates people with extreme views to publish on the web and then feed into spreading such topics.
If you look for confirmation, then often facts become less important. It is enough that you can easily find 10 or more people who claim something. I think this works equally good for all kind of conspiracy theories. No matter how idiotic the theory (“Chem Trails“), if you search for it you get hundreds our thousands of sources confirming what you were thinking.
The link to investing (finally…)
I think in investing, something similar is at work especially in one area: Believing that the end is near for capital markets. If you start to believe in this idea and start searching the net, you will find all sorts of proof that it is only a matter of days until the stock market implodes and the S&P 500 will drop by 80% or more.
Trump, Brexit or the Renzi Referendum are just part of the script which will inevitably lead to chaos and doom (not mentioning the 1000 conspiracy theories)
Clearly this could happen (with a low probability), but in my opinion it is both, very unlikely and clearly not carved in stone.
The stock market and the economy are very complex system and not even the smartest guy on the world can predict what is happening.
Interestingly, even I (although I hardly use Facebook a lot) constantly see ads or Emails which tell me “The Euro will collapse, act now and buy Gold” or “Protect your money from the coming Armageddon by reading this free report” and so on. I guess the reason is that Google knows that I am reading a lot about investments and therefore I get targeted by those who have a stake in a potential apocalypse.
The potential solution (for investors)
I have no clue how to stop the “Confirmation Bias Machine” in politics. Maybe with aggressive targeting by the good guys ? Or will the younger generations be smarter in processing fake news on Facebook ? Or will attnetion for Social Media will go down over time ?
But for investing I would suggest a very simple idea which, not surprisingly, comes from Charlie Munger (and was borrowed from other famous people):
Invert, always invert.
So if someone for instance makes a convincing case that Italy will go bankrupt soon, instead of searching Google for “Italy bankrupt”, one should try to dig out the positive aspects like that for instance the interest expense of the Government has never been lower and that things are happening in the economy and then make an “informed judgment”.
Or if other people argue that based on the cyclical adjusted CAPE based on normalized margins over 200 years, the S&P 500 has to fall at least by -50%, one should try to think: What do current companies like Google or Amazon have in common with a local Railroad 150 years ago and why should they earn the same kind of margins ? Maybe there is a reason that margins are higher than in the past with highly capital intensive companies ?
However an investor should also apply this technique when she/he is euphoric about something. Yes, Google is growing by 20% per year but couldn’t there be a scenario when growth stops really quickly, for instance when they have the majority of marketing dollars already in their pocket or they become so pwerful that the Government will reign in ?
Confirmation bias in my experience is one of the most common biases in investing and is often the cause for pro-cyclical behaviour and bubbles as well as significant under valuations. However it is also one of those biases which can be relatively easy be neutralized. For stock pickers, a simple Pro & Con list at a very early stage works quite well. If you don’t find any “cons” then you can be sure that your analysis didn’t go deep enough.
P.S.: This is an investment blog. So I will not tolerate any offensive comments etc. about politics or election outcomes.
Good post memy.
“Fake media” also in investing is a huge problem. The problem is what media to trust as we do not have the time to go back to the original data…..
On Cambridge Analytica: Good marketing, indeed. And since they cannot proof it, they are heroes to the ones who believe it…
On confirmation bias: I find myself ‘value investor confirmation biased’ all the time 🙂 For example when reading about technical analysis and their chart patterns I almost always have to think ‘Are you serious, dude?’.
Howard Marks’ “Second Level Thinking” might be another good example for how to treat with these kind of situations. Similar to the invert-classic.
By the way I think the ‘everything will collapse’-movement is becoming bigger and bigger. Whenever I talk to someone about investing 90 % are convinced that real estate, gold and Tesla (EVs are the future :-)) are the only safe harbours in our world today…
Der Tagesanzeiger, da brauch ich gar nicht weiter lesen. Wenn Schweizer Medien in Deutschland zitiert werden, würde ich vorschlagen, sich erstmal mit eben diesen zu beschäftigen. Ich habe auch ein paar Jahre in der Schweiz gebraucht, um zu realisieren, wie dir Medien hier gepolt sind. Sieht auf den ersten Blick seriöser aus, als es ist.
the story also was in Spiegel and other German speaking newspapers.
I also have lots of prejudices against the “Tagi” very populist. NZZ is much better, despite their conservative views (another biasing!)
thank. I aggree that NZZ is a very good newspaper…
Re the “Confirmation Bias Machine in politics” I guess that it’s cyclical like everything else in this world; when ultra-targeting is the new normal and political platforms become too complex I guess someone will go the other way towards more “old fashioned” politics to stand out.
I have a similar feeling….