Have you ever second-guessed your phone or car’s GPS?
I have. Mostly to my detriment, but not always. I’ve done it when the traffic ahead of me on the route is just too heavy and I find it difficult to stay put. In these circumstances the alternative routes I have tried are even worse. My bias has been that I’m too impatient to just sit there. However, after a few of these errors I have learnt my lesson and now trust the GPS.
On the contrary, there are times when my judgement, due to previous experience, is better than the machine. In these instances the GPS directs me along a route which requires crossing a busy road. It doesn’t factor in how long it takes to cross this road. You just sit there waiting for a gap as cars come from both directions. The gaps are small, and you literally take your life in your hands as you shoot across the road. It’s worse and more dangerous when there’s a blind rise.
It seems the GPS system hasn’t yet factored this in when calculating the quickest route, but I have previous experience to know. No doubt the humans behind the machine will code this aspect into the systems in time, or better, the algorithms will learn this nuance by themselves from the data gathered as drivers use the system.
Human vs. Machine in the world of investing and asset management
If you do a Google search for “Human vs. Machine” it yields nearly 2 billion results. The age of the 4th Industrial Revolution (4IR) has firmly taken its grip. Artificial Intelligence (AI) and Machine Learning (ML) are at the centre of this revolution and nothing will be untouched by these technologies – not least the world of investing and asset management.
The word “Fintech” is the catchall phrase for any business or application applying 4th Industrial Revolution technologies to the existing businesses in financial services. Wikipedia says that “Fintech, is the technology and innovation that aims to compete with traditional financial methods in the delivery of financial services”. I’ve italicised the word “compete” because above all, the really big debate and fear among us, is to what extent the “machines” will put us humans out of our jobs.
Currently, this isn’t happening for the most part across the majority of our industry in South Africa. Asset management is still mostly human-powered and, we lag the world in terms of the number of existing and new businesses adopting 4IR technologies. In my opinion this is merely a time lag. We will get there by necessity! The big question is – which businesses will it be and which ones will be the casualties for being too far behind?
The typical model of innovation is that existing and large business don’t disrupt themselves. There are many reasons for this, but typically they’re too big and too slow to react before it is too late. There are many business school case studies that highlight real-life examples such as Kodak; Blockbuster Video etc. A quote by famous Greek philosopher also comes to mind – “It is impossible for a man to learn what he thinks he already knows.”
Responsible use of innovate technology in the investment industry
A question one may consider is this: in the management of other people’s money, should we be experimenting with new innovations before they are proven to be beneficial, and to not harm our customers? Perhaps it is better to wait for those who manage their own capital and can afford to take on the risk, to prove that such new technologies actually work?
My answer to this issue is that active asset managers take risks everyday with their clients’ capital, in the pursuit of “beating the market” notwithstanding that the evidence bears out – the majority of active managers can’t and don’t beat their benchmarks after fees. For the minority that are able to – is it Skill or Luck?
Some of the answers to this question can be found in the book “The Success Equation – Untangling Skill and Luck in Business; Sports and Investing” by Michael J. Mauboussin. In the book the author explains the inter-relationship between skill and luck, and demonstrates that the more skilful we are as a collective, the more that luck separates the winners from the losers. I doubt anyone will disagree that in our industry the level of skill is as high as it has ever been! To paraphrase the words of “Dirty Harry” – Do you feel lucky?
My point expressly with the above is that exploring innovative technology, if done responsibly through proper process and governance is no more risky than how clients’ assets are currently being managed. The upside is that those that do adopt 4IR technologies may well find themselves armed with advantages over their competition which they didn’t have before.
Investors who wish to select the best managers to manage their capital ought to be asking questions about the adoption of 4IR technology.
Where and how to start applying machine learning within asset management?
Starting with the lofty goal of using machine learning to help make economic forecasts; picking stocks or allocating assets are perhaps overly ambitious and will yield many failures before or even if it succeeds in a broad sense. Author and Wall Street Journal writer, Gregory Zuckerman reveals in his recently released book “The man who solved the market – How Jim Simons launched the quant revolution” – how it took decades for some of the best mathematicians (not investment people!) in the world to “solve” the market. They did however achieve their goal.
Of course it is not only in the areas of selecting securities or asset allocation where artificial intelligence and machine learning may provide benefit but in many other areas of the investing value chain – such as operations and client servicing. At a minimum investment businesses should be using 4IR technology in the area of business process automation. This can help control costs and make businesses more efficient and scalable – meaning fees can be lowered for the benefit of clients.
Given the power of these technologies there are 2 specific areas where I believe they will have massive benefit. Firstly and arguably the most valuable service our industry should provide is risk management. To quote Warren Buffet – Rule no.1 in investing is – Don’t lose money! Secondly, is the use of these technologies to help us mitigate our behavioural biases – many of the reasons active managers fail to outperform the market are due to behavioural tendencies. Through the application of things like data and natural language processing and, pattern recognition, human portfolio managers can improve their results if they embrace the tech revolution.
Vanguard founder, and father of index/passive investing, Jack Bogle, wrote his senior thesis making the case for indexing way back in 1951. It has taken nearly 70 years for the “global market” to fully recognize and embrace the validity and case for passive investing. I doubt the 4IR will take a fraction of that time to take hold, simply because the pressures facing active managers have never been greater and continue to rise.
My answer to the narrow question, Human vs. Machine? – is that Machine wins. However, the ultimate winners will be those businesses, new or old, that adopt a collaborative Human and Machine approach.