The pandemic-induced lockdown has unleashed an unprecedented level of uncertainty in the behaviours of Australians. In this exclusive interview, Dr Jones shares his thoughts on the Australian economy, recession, consumer behaviour, and socio-trends. Dr Jones and his colleagues, during this pandemic, have invested their skills, efforts and time in developing a forecasting model for predicting such disasters in advance giving the authorities ample time to come up with a coping strategy if not a prevention plan altogether.
Full Interview Transcript:
Dr Jones on his journey and how he founded Jevons global:
I founded Jevons Global in 2011, after a long career operating within large financial companies: funds management firms, stockbrokers, and investment banks. My intention in starting my own firm was to travel light and figure out where the future might lie given all the technological change that was going on. Prior to my entering finance in the mid-1990s, I had a first career in research, within Universities, Research Institutes and Defence Science organizations. The planning horizon of business is much shorter than that of R&D and yet there is benefit to thinking about the long term as well as the short term. In many ways, the journey has been more important than the destination this past decade as I wanted to figure out how to create a culture that could do both long term and short term thinking in the context of investment.
Some of the most significant projects he has worked on:
Data has always been a big part of my work, in finance, and much earlier in scientific research. One of the more significant projects I worked on was developing policy research for an organization called the Centre for International Finance and Regelation (CIFR). Starting in 2013 I was approached by the then CEO Prof. David Gallagher, to help frame policy for the Australian financial regulators regarding the use and importance of data in the finance industry. Along with the CIFR team, I led a deep dive on the state of data in Australian finance, which culminated in a report I titled RADAR. This was before the big trend towards open data for banking in finance. The activities of CIFR helped frame some of the background to subsequent policy and I contributed to that via various reports for the Productivity Commission.
In many ways, policy work is a game for later in life, when you have a broad understanding of where and why good ideas sometimes fail to find traction. The state of data in Australian finance is much better than it was when I did the work for CIFR. That is not because of us, but we helped change the climate.
Dr Jones on social media and data science inter-relation and how the data from social platforms is utilised:
Presently we use social media to track topics of wide discussion and importance. It does not figure directly in our investment process but that might change. One of the areas I focus on is the measurement of market sentiment. I have a method to do this directly from the data of what stocks traded each day and in what volume. This has been quite successful for me and I now hope to improve it further by combining that signal with what is also coming from social media. These are two very different data sets: one is simply numbers of price and volume traded, the other is text-based signals on what is trending and social reactions to that. In many ways, the future of data analytics is to combine more and different signals to gain further insight into what is driving trading decisions. If we had such a system operating in these last three months, we would probably have done better trading COVID-19.
Dr Jones on how the data is collected, stored, and cleansed at Jevons Global:
We use a number of lightweight python-based robots to collect data, either directly from an Application Programming Interface (API), which is like a programmable fire hose, or through web-scraping or direct download from websites. A good example of the latter is the Securities Exchange Commission (SEC) website in the USA, called EDGAR. EDGAR has all the corporate filings from listed companies. It is a very comprehensive longitudinal data source, going all the way back to the late 1990s. This dataset is really good for experimenting with new data extraction and processing technologies. For example, the data is in text form but contains embedded financial tables. Over time it gets easier and easier to extract such information using new AI tools designed for language processing. Earlier I mentioned my interest in combining the long-term perspective with the short term. Gathering, cleaning and extracting data is a perfect example of this focus, as it is a challenging problem where you can keep improving. Whenever we apply such technology to ancient history (in finance that is what happened yesterday) I say that I am doing Financial Archaeology. That makes it sound cool, but in fact, the data cleaning applications of AI are cool. You don’t want to process this stuff by hand. There is way too much of it. However, sometimes I like to dream about what it would be like to have a robot good enough to read the old Sumerian clay tablets describing what happened in their market in their day, three thousand years ago. The simple truth is that human nature never really changes, but how we record what we do does. This is where modern AI is really powerful. You can think of it like a telescope, it can reveal detail in the far-off distant past, or bring the present action up close and personal.
How does data science add value to the business? Is it focused to drive revenue, reduce costs, and increase product and service for the organisations, or is there more to it?
Up to recently, our main use of data science has been to find an edge. I mentioned that I have spent a long time in the industry working in large firms. These often had very sophisticated systems for trading automatically along with the more traditional stock selection models for risk and return. I know how those models are built, so much of that, while impressive, is old news to me. The challenge is to rethink how we collect our data, process that and model behaviour in markets. We feel we have made great progress on the efficiency and cost side, and now the challenge is to create value in the business by building revenue. My aim has always been to get the early part right, the tedious data stuff right, so that one could attract talent that really wanted to do the exciting stuff: build new models and strategies to win in the financial markets. There are so very many great people in this game that it is a hard game to play and to win without putting in the groundwork. I am always mindful of a comment the legendary Jim Simons (of Renaissance Technologies fame) made when asked what helped drive their success. He said that they spent a lot of time and energy early getting their data right. I have followed that credo.
Dr Jones remarks on how Data Science/Predictive Intelligence could have done better in preventing the spread of COVID-19 from China to global hotspots? And considering we had SARS, was it possible to predict the impacts of COVID-19 before it started?
I and a good friend of mine, Stephen Duchesne, of Castellan Capital Pty Ltd, had put a lot of effort into the problem in February when the epidemic was pre-pandemic. We found that we could build a very good forecasting model by adapting ideas from the modelling of interest rates. That was helpful to us but did not gain any traction with the wider community. What we found was that a good digital marketing effort to garner attention trumped a great model. That was a bit disappointing to us that we had a great model that was barely used when it mattered, but it was also an important lesson. These days you need very well-developed channels to get your message out. There is such a huge volume of messaging that the quality of the message seems immaterial to the impact. I am working to fix that.
Taking the Australian economy as a whole, how the economy is tracking at the moment?
This is the Twilight Zone right now. We are not sure how deep or long the recession will be. What we do know is that some industries, such as tourism and hospitality, are really copping it. Digital has been strong, but not just the new entrant companies. Kogan is doing great, sure, but so is Coles and Woolworths online and also the various businesses within Wesfarmers such as Catch. I think the future is hybrid. When you spend long enough in financial markets you will one day wake up as an Old Dog. No Old Dog ever likes to think that they cannot learn new tricks. I think the future of digital is very positive for businesses of all types and sizes. I just bought a new apartment in Canberra and so am shopping for furniture online. Some of that might be contemporary design, some might be classic like Scandinavian Modern from the 1950s or 1960s. It does not matter. I can find both old and new online. When you move, after twenty years in one place, you find you own a lot of stuff you don’t want to. The solution is to sell that online. Habits change. I thought Amazon was cool in 1998, but I very rarely buy stuff there now. These days I am more likely to sell something on Amazon than to buy something on Amazon.
People are discussing the recession these days. How deep would the recession be? Do you have any data science models to predict it?
Predicting recessions is difficult and in many ways not so relevant to financial trading. The markets will often move ahead of the real economy. They seem to have done this pretty well this time around. Like many experienced markets folks, we were very sceptical on the early rally in the USA. However, on the recent employment data it does seem as though people are getting back to work. I think what this latest episode will do is awaken interest in high-frequency data sources such as transactional data. It would seem that folks who were onto that earlier did better. That is good to know. Data is data. Once you know that a piece of data has value you will incorporate it into your model. Survivors in the financial market game are like Borg, from the show Star Trek. You might take a few hits, but then you adapt.
Will the economic crisis caused by COVID-19 changes some industries or people’s consumer behaviour forever?
I think it will probably lead to a new set of, shall we say, genteel habits: respect for personal space and standing a little more distant on first meeting. I do not think that will become as elaborate as what they do in Japan when greeting one another, but some change there seems guaranteed. On other things, I think online grocery shopping is one huge change. For me, I moved cities. I figured that I could go live in a place like Canberra, which I like for a peaceful lifestyle, and not have to bother about the fact that I was not in “the centre of things”. People will tire of Zoom meetings, but they know how to do them.
What are the new socio-economic trends do you see?
Generational change is always the one that sneaks up on people. We don’t really know what the post Millennial generation thing will look like. Whatever it is will be guaranteed to shock Millennials. That seems to be the way with socio-economic trends. Whoever felt settled is in the group to be shocked. On a more serious note, I think that Climate Change is taken more seriously now, and that will likely dovetail with practical decisions on where to live and how to work. Oddly enough, the new workplace of today looks a lot like spaces I have used in the past: my mother’s Art Studio or my old professor’s Lab.
Forecasters at the reserve bank of Australia and treasury are relying on an economic model which assumes the deeper the recession the faster recovery we’ll have. Based on the global trends and depth of your knowledge, what can you predict, will this model work?
I think things will be less bad than first expected in terms of a rebound in most consumer activity. However, the recovery will likely be slower than expected in some areas like tourism. It is hard to predict so I won’t try, but I think airlines are a very tough place to be. Automobiles might surprise the other way. We just don’t know, but some people might decide to buy their first car. It might well be electric. Consumer behaviour change is always hard to predict, but we can try and recognize it when it happens. At some stages of life, when you are most settled, it is hard to recognize change. At other times, it is all of a sudden easier. At this time, for me, it is easier to spot changes. That situation has varied a lot through my life, so, when the time comes. you have to leap on the opportunities you can actually see.
Previously the Australian economy was heavily reliant on mining and manufacturing. Still, in the recent decade, Australia is growing in other industries such as healthcare, technology, and construction, according to you what other sectors would be booming post-COVID-19 crises. Will our economy continue to diversify?
I think Australia will continue to diversify in healthy ways. We have a much better-educated population than when I grew up in the early 1970s. However, at that time we also had pretty sophisticated labs for Research and Development. A lot of that got wiped out in the 1980s and post-Cold War rationalization. Now we have to build up this capability once more. It can happen through the renewal of old industries like mining, through automation, new industries like digital services and retail, and frontier areas like the emerging and nascent Australian space industry with all the promise of advanced manufacturing like 3D printing of rocket nozzles, and all the fun stuff. There is tremendous opportunity ahead for those who are willing to grasp it.
Would you like to add any last words, about the future of Australia?
The big positive I see from COVID-19 is an opportunity to break with the Tyranny of the Two, meaning the two mega-cities of Sydney and Melbourne dominating economic activity and conversation. Now that we have seen that remote working does work, and people can do Zoom, even if too much of it drives them a bit mental, then we can increasingly grow our economy from any place. Australia is so blessed with so many great places to live, that I think we in business should embrace the opportunity. This is why I moved to Canberra, the place where I first studied at University back in 1980. I love the place because it is what you make of it. Nobody ever got famous for living in Canberra. This means people just get on with whatever they are good at. This town is home to Australia’s top-rated University, my old alma mater, the Australian National University (ANU) and is the de-facto home of our Space Program. I came here to do some finance, do some mathematics, make some business and walk in the forest.