By Mrugank Paranjape
With the development of transparent markets for trading multiple asset classes from commodities and bonds to stocks, the data that gets generated in the process has immense value in itself for decision-making. It can be used in varied activities from mining and production to manufacturing, alas apart from trading decisions! Rightly so, the British mathematician Clive Humby (2006) projected that “Data is the new Oil.” Data available from and for the commodity markets today continues to grow at an unprecedented pace, thanks to the increasing trend of digitisation of commodities value chain and the transparency that it enables. Most of such data still remains unstructured and unharnessed. Refinement of this data gathered so from the value chain through use of data science and its use can be a game-changer in today’s cutting-edge competitive commodity markets.
According to an estimate of the government of Singapore, about $10 trillion worth of commodities are produced and consumed per annum globally. With each commodity comes its own set of challenges from the way it is produced/mined, refined/processed, traded, marketed and managed in terms of digitisation and transparency of the value chain. On top of that, driven by supply-demand fundamentals along with geopolitics and several other factors, volatility is the only constant in the cyclical commodity markets and so are the risks associated with business decisions of the stakeholders. For commodity stakeholders who generally operate on thin margins, situations of price instability make risk management inevitable. The multi-trillion-dollar global commodity trading industry trades financial contracts with underlying commodities that are crucial inputs for much of the manufacturing sector, such as crude oil, copper, cotton, etc. These markets not only provide advance signals, but also facilitate risk management by them, and thereby play a vital role in greasing the wheels of the economy.
For investment banks that support healthy existence of such financial markets in commodities, back-office work is vital to efficient management of theirs and that of their customer positions in the markets. Gone are the days when back-office work for position management in investment banks involved large capital expenditures and a long time period. Position management can now be done seamlessly across the life cycle of a traded commodity position with the help of agile technologies like big data, blockchain, machine learning, artificial intelligence (AI), robotics, etc, which provide for efficient estimate of demand and price swings. It is, therefore, imperative that we are aware of these technologies and know how they can impact the world of commodities.
In simple terms, big data is literally a big or massive amount of raw, unstructured and unformatted data that updates constantly to uncover patterns and relationships, thereby empowering decision makers to make the right decisions based on real-time analytics-based insights. Big data can be defined by its volume, variety and velocity of its collection. Its harnessing is likely to be the next frontier in technology essential for competition and efficiency. Here, ‘data’ can mean anything from structured databases to written data, text, photos and videos, which would need specialised software for converting into utilisable data points for the purpose of decision making.
Price movements, changes in market cycles, new regulatory frameworks, etc, create millions of individual data points that can be processed to provide efficient inputs into decision making not only for back-office managers, but also for the markets and may even provide sufficient policy inputs at the commodity-economy level. Besides, efficient use of these data can provide feedback on market conduct and help commodity traders in making efficient decisions about entry or exit points. Wide-scale adoption of big data can provide for competitive businesses in the economy and hence the competitiveness of the economy itself.
An important contribution that comes from the creation of Bitcoin is the distributed ledger called the ‘blockchain’, which, in simple terms, is a distributed database shared across a defined network. Each computer across this network has a copy of this database and every bit of information is mathematically encrypted and named a ‘block’, and a chain of such ‘blocks’ not only validates transactions, but also the storage of the underlying asset. A World Economic Forum survey (2015) suggested that 10% of global GDP worth of economic transactions will be stored using blockchain technology by 2027, including the trail of transactions leading up to their current ownership.
For commodities with a physical dimension and quality parameters that help price them, adoption of blockchain in transactions and storage will enhance efficiency in execution of commodity transactions and storage with the associated information set. Blockchain has the ability to bring all the stakeholders of the commodity market together to prevent fraud, eliminate third-party, speed up clearing, thereby improving transactional efficiency and financialisation while bringing in operational efficiency in the value chain. Further, blockchain-based transactions will enhance regulatory filings and reporting by improving market transparency and auditability. The fact that the commodities industry has woken up to this potential has been visible in attempts to deploy blockchain across commodities/verticals such as power, diamonds, food and oil. Large trading houses such as Gunvor, Koch, Trafigura and Mercuria have started trials using blockchain technology for settling their back-office trades.
AI and machine learning
Machine learning or AI, by definition, is the implementation of computer software that can learn autonomously. Machine learning and AI can offer new opportunities to improve process performance and realise significant cost savings to market stakeholders. To achieve the final objective, AI/machine learning uses the structured big data and learnings by linking patterns with the fundamentals and price movements with appropriate level of noise reduction and normalisation, thus improving the decision-making process and thereby improving efficiency across commodity corporates that use AI/machine learning. A predetermined logic based on AI/machine learning will allow traders take instantaneous decisions on commodity curves and settle trades, and therefore enhance transactional efficiency in the markets.
The way forward
With the help of the emerging trend of digitisation of commodity transactions and storage, industry can take a big leap forward if the same data can be appropriately collected and harnessed using big data and AI/machine learning. Increasing adoption of blockchain in storage and transaction of physical commodities would not only enhance transactional efficiency in the markets, but also generate adequate data to be used for corroboration with other big data to help AI/machine learning to further efficiency in market-based financial transactions. As each of these technologies are interdependent on each other to bring about holistic transformation in the underlying markets, it is essential to have a policy regime that not only financially supports developers and users of the technology, but also provides supporting policy environment in terms of its implementation. Further, public institutions in commodity storages would also be actively encouraged to take advantage of digitisation, and also provide for enhanced transparency for efficient decision making in the markets. As in society, harnessing strengths of digitisation of commodities will help us move on the path to becoming ‘price-setters’ in the global markets.