The Power of Language

“Language” has featured heavily in the UK media coverage these past few days. The events in and around the UK parliament since it became non-prorogued were strangely both shocking and entirely predictable which sounds oxymoronic. Indeed that word without the “oxy” part was being widely used to describe our elected representatives before the (non) prorogation started.

Although I’ve sadly become accustomed now to polarised exchanges along the lines of “I'm right and you’re an idiot” I still optimistically yearn and hope for a return to respectful intelligent centre ground debate. I look for evidence from multiple perspectives , listening and compromise to return to the big issues of our times. It seems I may have to be very patient because that is like swimming against the tide presently.

On all big and complex issues the tone and language used in Parliament seemed to hit new lows reflecting perhaps the kind of exchanges we have grown accustomed to seeing on social media which largely exists to encourage and amplify an unfiltered expression of views , not necessarily based on any particular insights or knowledge. This is the wisdom of the crowd or as some others may suggest the madness of the crowd.

So the recent newsflow got me thinking and reflecting about the power of language and my own personal journey dabbling in behavioural finance. It helped me make the possibly tenuous leap that is the subject of this thought piece. This is actually about about a lesser known but fast rising part of active fund management which (among other feeds) aims to harvest what folk say about products and companies on social media as a predictive signal of share price movements. Each word and phrase assigned a score in real time. I am talking about the use of so called “Alternative Data” using machine learning techniques – most specifically Natural Learning Processing (NLP). Its a really big business and is considerably more advanced amongst quantitatively minded US fund managers than our UK counterparts to date but I expect the gap to close. 

Most folk intuitively know that the growth of data is following an exponentially rising path , but when you stop to spend time with data scientists and relevant specialists to see what that means in charts and statistics it is truly mind blowing. The game of active fund management is getting harder – most of those in the know would say this rising tsunami of data and data capture is making markets more efficient . So the edge required to find “alpha” is to be found in so called “alternative data”. The jury is out still about whether it works because without doubt this data includes a lot of bad data as well as good data. And data mining is an ever present risk. 1991 Nobel prize winning economist Ronald Coase famously said “If you torture data long enough it will confess to anything”. 

Well today there is a massive industry in alternative data capture , sorting and connection to find various edges or insights. Indeed it could be said that there is no longer any alpha in traditional data ( I expect some push back on that view in particular). But I’m convinced the fund management industry is entering a highly disruptive phase that is long overdue and now we can see some of the instruments of that disruption. This is exciting and it presents opportunity. The fact that ESG data capture is in its infancy is but an example of such opportunity. It’s become easier for us to see emerging winners in other industries undergoing disruption – eg media or transport etc but services can be similar and that means winners and losers. New “information asymmetry” arising from new ways of processing new forms of data could be where the advantage lies. (And I don’t just mean language but that is the hook for this article).

To be clear these are entirely my personal views in this article but I’m conscious of representing a brand too. So I ask forgiveness (especially from my employer) as I return to a political example that partly illustrates my personal learning curve and just how much further I have to travel on it! This news story breaking on Thursday night and Friday morning had me ranting at the TV (OK I still use mostly old school sources of data and at least I wasn’t ranting on social media!) . A nice weekend away since clearly hasn’t quite expunged that feeling however. 

 https://www.bbc.co.uk/news/uk-politics-49854569

 https://www.bbc.co.uk/news/uk-politics-49847304

 So I had just seen the footage on the BBC news of Dominic Cummings being confronted by labour MP Karl Turner who appeared to be talking with real emotion about the consequences of damaging language in our parliament. Cummings reaction didn’t surprise me. Soon after Amber Rudd weighed in on the idea that the actual words used matter (I agree) and some can incite violence. (well to the clever data scientists some words and phrases will also point to making money). In Cummings' case he really knows the power of language and in choosing exactly the right words and phrases to move opinion. I'd seen that evidence before.

I first saw Dominic Cummings in person at an event called “Nudgestock” held in Folkestone in 2017. Most of his speech and Q&A at the event is available in this link here (although my biased and faulty memory suggests a couple of pearls are missing or redacted from the video) 

 https://m.youtube.com/watch?v=CDbRxH9Kiy4

I recall vividly that he totally sent shivers down my spine at the time as I listened but I was also enthralled in equal measure by what he said. A weird experience which hasn’t left me since. It was also my eye opener for “psychographic profiling” which with conscious irony I posted about on my personal Facebook account at the time referencing Cambridge Analytica. (who became big news the following year).

I attended that NudgeStock conference because of my fascination with the potential of embedding more Behavioural Finance awareness in our consulting work. I was struck by how few of the asset management and consulting industries were present at the time unlike other industry sectors (especially advertising) and academia. Mind you Folkestone is a bit off the beaten path. 

So wind the clock forward over 2 years and new applications of this data science is becoming a key part of the money management techniques behind some members pensions savings. Of course such smart fund managers were always way ahead of my thinking but they just didn’t do coastal visits perhaps. Is Cummings strategy of pitching Parliament / Establishment versus the People winning? It looks like he thinks it is and I hear evidence of it on the street and on the radio chat shows etc. Are the quantitative (and tech savvy fundamental) asset managers harvesting a new repeatable alpha? The jury is out... but I think it’s looking promising.

Folk who know me , know that I’m a keen advocate for developing Responsible Investing programmes for institutional asset owners. Although in UK/EU at least the Regulators are on side , the public debate is better known and the academics are aligned, we still face a lot of barriers around the quality , consistency and comparability of ESG data. This is perhaps one area where the new highly brainy breed of alternative data specialists in the fund management industry can help. While the behavioural and cultural aspects in our Parliament seem to be going backwards (on all sides of the House) I have absolute faith that in commercial business the culture and values are getting better and better. But where's the proof? Like many , I’ve commented before about the social purpose of business but its just opinion. Perhaps NLP (Natural Language Processing) will help analysts capture and harvest the cultural aspects of the businesses we (trustees , asset owners) invest in , consistently and comparably. This could be “good” data.

And finally what would NLP make of what’s being said in Parliament presently?! I’d like to think it would be discarded as “bad data” – at the very least it was bad language.

The power of language....

John Belgrove 29/09/2019