A hosted white paper by Etienne Amic, veteran energy trader and founding partner of CommodiTech Ventures.
It seems almost impossible to escape headlines about robots taking human jobs, blockchain making the financial system obsolete, quantum information delivering a kick to Moore’s Law or the FAANGs’ meteoric rise toward trillion-dollar market caps. Is this technological frenzy set to affect the commodity trading industry, or are there solid reasons why this cannot be the case?
A recent study by Boston Consulting Group described the relative degrees of digital adoption across eight industries. It should come as no surprise that energy commodities ranked last and the media ranked first. But the change that information technology is starting to drive in commodity trading is becoming visible, and the rate of digitalisation will only accelerate. There are key areas in which technology is affecting the industry, as described below.
Physical fundamentals data. The amount and types of data available to traders – from the specific grade of crude on any tanker globally to real-time temperature measurements from over 200,000 weather stations worldwide – have exploded in recent years. Advancements in satellite technology and remote sensing have narrowed the sampling frequency of most parts of the Earth to a few hours. As an illustration of the shrinking communications gap, Planet Labs put 88 of its ‘doves’ (each weighing just 4kg) into orbit in February 2017, followed by another 48 in July. The combination of reusable launchers and mobile phone technology finding its way into much lighter satellites allows vessels to be tracked and inland inventories to be assessed as never before. The trader’s edge is no longer ‘boots on the ground’ but ‘eyes in the sky’, and this is contributing to a commercial space race, with start-ups like Spire, DigitalGlobe and exactEarth in the running.
Open-data initiatives by governments – such as the release of customs data or regulations forcing the timely publication of power plant dispatch and flows in gas pipelines – have shed light into previously shadowy corners of the market. The Internet of Things (IoT) and a new slate of government-mandated initiatives such as smart electricity meters are the next steps in this vast movement of digitalisation. It is likely that within five years the supply and demand of oil will be visible in real-time to an extent that will make us question why initiatives like the Joint Oil Data Initiative were launched.
Analytics. While huge streams of data are appearing at an accelerating rate, the cloud provides the natural infrastructure to gather them at scale and process them with negligible delay. Specialised hardware such as Nvidia’s GPUs or Google’s TPUs, coupled with deep learning algorithms, have contributed to the transformation of other industries almost overnight. Of course, this engineering must be backed by platforms providing a logistics service and a viable economic model, but the bulk of value creation is in analytics. Computer vision, speech recognition, natural language processing, text-to-speech conversion, game playing, ad targeting and search have been revolutionised by the trinity of cloud computing, specialised hardware and machine-learning algorithms. Make no mistake – this collective force is coming to our industry, and it is going to be an arms race. Adoption has taken longer for us is because we are dealing with real-world data, which is www.humancapital.global 2
inherently noisy and only partially digitalised, while the Internet economy has 100pc coverage of very pure click data. Start-ups like Orbital Insight, Rezatec, TellusLabs, Kayrros, Ursa and Vortexa are racing to provide better analytics to commodity traders.
As the breadth of data sources expands and we acquire more history, algorithms will get better at making connections between fundamentals and price action and offering logistical optimisations. In the transition period and probably beyond, the challenge will be to build teams combining expertise in commodity markets, data engineering and data science and to decide what is best done in-house and what is best delegated to start-ups. This is easier said than done.
Sales. Given the complexities of the supply chain and the physical nature of the business, commodity trading has always been a highly intermediated industry, with large swathes of it being dealt over the counter (OTC) and on the phone. But the traditional middlemen – brokers and salespeople – have fallen in number as more markets trade electronically and technology enables direct market access (DMA) for clients. Hedge fund salespeople often become sellers of software solutions (prime brokerage) rather than dealers of derivatives. This trend is sure to continue – there are entire commodity markets that would benefit from DMA.
Trading. Gone are the days of trading by instinct. While equity and foreign exchange markets have been dominated by electronic trading for more than a decade, algorithmic commodities trading has only started to take hold. There is a lot of confusion about algorithmic trading in our industry. Are we talking about: the micro-slicing and execution of macro orders decided by human traders (who seek to hide their intentions); the ‘sniffing’ algorithms that monitor order books across multiple venues to detect and front-run large orders; programmes that process market news and fundamental data and send orders to trading venues in reaction; or algorithms that read the liquidity available in order books with a view to trading the optionality of physical assets? These are all very different beasts.
Traditional traders have tended to lump these algorithmic plays together as nuisances that operate on very short timescales and are irrelevant, if annoying. This is an error of perspective. What these separate algorithms are trying to emulate are the distinct skills deployed by human traders in their day-to-day jobs. Objectives could include assessing the liquidity available to place a large order, reading news and reacting to it in context or delta hedging a gas pipeline modelled as an option on the spread between two hubs. Many traders would benefit from a more proactive approach to algorithmic trading – the chess world champion is neither a program nor a human but the combination of human and machine. All trading venues, whether OTC or exchange, have developed trading APIs with fast access. A good way to know if algorithmic trading is dominant in your market is to check whether the exchange (or MTF) has imposed financial penalties on parties with a high ratio of orders to transactions. If that is the case, it was in response to the rise of algorithmic activity. Human traders should take heed – however good you may be, the evolutionary forces of the market will prevail over time unless you start adapting. www.humancapital.global 3
Systems and operations. The commodity industry is infamous for its inefficient and tedious workflows – paper, spreadsheets, email and phone usage abound in middle-and back-office processes. This is not to say that all commodity chains have stood still. In European gas and power one can execute, confirm and schedule standard physical trades without much human intervention at all. The situation is clearly lagging in the case of physical oil or metals concentrates – the complexity of transactions (involving four or five parties) is a poor excuse for the lack of progress. But the sector has started to upgrade its archaic functionality with digital solutions. Aquilon’s Energy Settlement Network in the US or Autilla’s attempt at modernising precious metals back-office and reporting are good examples of this trend.
Even in risk management, where technology has played a more central role, users are shifting away from time-consuming in-house builds or bulky third-party CTRM implementations to more lightweight, flexible and cloud-based solutions. Beacon and Molecule are two examples of start-ups pushing into that space. There are significant savings to be made from software whose improvements are available to all users instantaneously.
Assets. Managing fleets of power plants has historically been resource-intensive and time-consuming. For instance, trading teams of seven or eight working day and night shifts would be required to ‘optimise’ 2-3GW of wind and solar assets. Real-time recommendation algorithms offered by companies including Likron now remove the need for large teams and produce more accurate results at much greater speeds. Coupled with IoT, larger assets are now firmly in sight for these algorithms. If autonomous cars are inevitable, self-trading power plants are not far behind.
Marketplaces. Becoming the next Amazon marketplace is the industry incumbent’s dream. Incredibly, a market as large and significant as physical crude oil is still mostly conducted offline through brokers, over the phone and instant messaging. A global electronic marketplace – where prices are displayed on-screen and buyers and sellers can execute trades directly with the click of a button – is a seemingly obvious proposition, but it does not exist. There are a handful of start-ups striving to create the go-to global trading platform for their respective markets, be it TradeCloud for refined metals, BunkerEx for bunker fuel or OpenMineral for concentrates. It is obvious that widespread industry adoption will be the key determinant of success. Even clearer, though, is that success is far from guaranteed – some ideas can be logical but out of step with the industry’s natural evolutionary timescale.
People. Recruitment has also been affected by technology. Traditionally, front-office roles were secured by finance or engineering graduates with a highly commercial leaning. The sought-after profile is now that of a data scientist or computer programmer. Companies are also no longer looking to hire older-generation star traders – they prefer to raise young, highly technical analysts who are well-versed in Python or C++ through the ranks.
Not optimised for success
I do not advocate technology adoption for its own sake. In fact, technology is often derided for offering solutions that do not address real-world problems. But the degree of innovation and the pace of adoption of new technologies in commodity trading has been particularly slow. Why is this? www.humancapital.global 4
Structural complexities. Innovation in the commodities B2B environment is challenging, with long supply chains, multiple stakeholders and differing physical attributes (eg grades of crude or types of coal) to contend with. It is hard to imagine the speed and extent of the ‘Uberisation’ that upended the taxi cab industry being replicated in wholesale commodity markets. For one, the commodity trading industry’s level of service to clients has generally been good, whereas taxi drivers have historically been unpopular in most large cities. But this is not enough to explain the wide gap that has opened between B2C – where the consumer can shift its purchasing behaviour with the swipe of a finger – and commodities B2B. Would you rather have your copper handled, stored and delivered by Amazon, or by the incumbent players? It is painfully clear that not enough has been invested in information systems over the past 20 years. On the other hand, the seamless buying experience we know as consumers has been made possible by huge capital outlays over sustained periods. Critical parts of the commodity supply chain – logistics and warehousing, for example – are being encroached upon by ambitious young companies such as Transfix and Flexe. This trend is only going to accelerate – capital will find its way with solutions more in line with today’s standards.
Lack of expertise. The extent of innovation that we see in other industries including the media and retail stem from the fact that those sectors are already well-advanced in their digitalisation cycle. This self-reinforcing phenomenon means that those sectors attract the best technical minds and, consequently, produce the most innovative technologies. Innovation has become the norm in these industries – it is written into their DNA – with the CTO or CIO often playing a more crucial role than the CFO or COO. This is simply not the case in commodities.
Culture. The factor that has perhaps the greatest effect on the level of innovation in the industry is behavioural. Commodity trading has not been very welcoming of change. Much of what technology has come to represent – transparency, efficiency, democracy, open access – is anathema to an industry that has been built on secrecy and the competitive exploitation of market inefficiencies.
Learn from history
All industries are undergoing profound change, and much of it is being driven by technology. We would do well to take a step back, take a good look around (inside and out) and try to make sense of it all. What lessons does history offer us?
Technology trumps tradition. History is replete with examples of technology-focused enterprises outgunning their more traditional counterparts. A prime example of this was the Intercontinental Exchange buying the centuries-old institution that was the New York Stock Exchange. It comes as no surprise then that the top-five companies as measured by market cap are all technology companies. These spots were filled by energy and financial services companies only 10 years ago.
Its impact is pervasive and indiscriminate. Examples of incumbency and market share not guaranteeing survival abound in modern history. Borders – once the biggest book retailer – decided to outsource its ‘non-core’ online sales to a company called Amazon. Blockbuster famously rebuffed an opportunity to acquire Netflix for $50mn. There is also the hard lesson provided by Kodak, which dismissed the threat of the digital camera for fear of cannibalising its core business and is now a new start-up willing to dabble in the blockchain to protect the property rights of images. www.humancapital.global 5
Do not underestimate the threat from outsiders. The complexities of the commodities industry may so far have been a barrier to entry, but we cannot overlook the impact Airbnb has had on the hospitality sector, or Amazon on retail. Closer to home, former NRG Energy CEO David Crane once wrote that “There is no Amazon, Apple, Facebook or Google in the American energy industry today”. He should really have added “yet”. Google and Apple are already fully licensed wholesale power traders in the US, and I would not bet against these ‘outsiders’ if or when they decide to really compete with established traders.
What can we do better?
If there is one industry that has heeded the lessons outlined above, it is financial services. Over the last five years or so, established players have realised the futility of trying to maintain the status quo or compete with nimbler digital start-ups. Instead they have embraced the opportunities for innovation that FinTech brings and have taken a much more open-minded approach to adopting new technologies, creating innovation labs and incubators along the way.
The commodities trading industry would be well advised to embrace its own ‘CommodiTech’ movement. Specifically, this can be achieved through the following actions.
Collaborating. Partner with technology start-ups via strategic partnerships. Even if the solutions offered are not 100pc perfect or fit-for-purpose, incumbents should experiment with these new tools and offer market feedback to help improve them. The costs of this approach are typically lower than pursuing such projects in-house.
Investing. Access new and emerging technologies by funding start-ups through early stage venture capital. For example, some banks – and notably Goldman Sachs – have made ‘strategic investing’ a business in its own right and have collected solid profits along the way. The key in the commodities trading industry is not to put off the investee’s clients – which may be competitors – by potentially investing through an independent fund.
Evolving culturally. By far the most difficult – yet arguably most important – success factor is adapting the mindset. Technology has historically been seen as a support function, with its impact on profitability limited to cost-efficiency gains. Business leaders need to treat technology as part of their commercial offering.
Taking a practical approach. From the entrepreneur’s perspective, start-ups seeking to enter the nascent CommodiTech space should take a practical approach to innovation. It is naïve to think that one can barge into an industry as large and long-standing as commodities with claims of ‘disruption’ or ‘revolution’ and expect to be welcomed with open arms. Young entrepreneurs are better served by reinforcing their advanced technological know-how with solid industry expertise by partnering with experienced commodity traders or managers and, similarly, forging partnerships with industry leaders to enhance new digital tools and solutions. www.humancapital.global 6
The profound impact that technology has on industries is more apparent now than ever, and commodity market participants face a choice – innovate or risk obsolescence. The industry has been in place for centuries, being one of the most complex in the world and critical to contemporary life. We must respect its systems and processes, but to ignore the profound and potentially transformational forces of technology is to endanger its legacy. Business leaders must embrace innovation now if they are to survive – let alone thrive – in the new age of Commodities 2.0.
Etienne Amic has more than 20 years of experience as a commodities trader, investor and entrepreneur. He is founding partner of CommodiTech Ventures, the first specialised early-stage venture fund investing exclusively in commodities-related technology. He is also chairman of leading oil data and analytics company Vortexa and EnAlgo, a power and gas trading technology company. Amic was head of European energy trading at J.P. Morgan and Mercuria and began his career trading crude oil at TOTAL.
View the PDF version here.