The recent plunge in crude oil price has been making a lot of market noise these days, but what caught me in is not about the increasing supply from U.S. shale nor weakening demand due to global economic slowdown. It was the news which talked about how some hedge fund managers made money from current decline in oil price, by using artificial intelligence (AI) to scan and analyse satellite images (e.g. looking at the shadows cast by floating-roof storage tanks) in order to help estimating the change in global crude inventories on a more real-time basis. Although this kind of satellite imagery intelligence is nothing really “new” to the market, the accuracy and reliability of these analysis continues to improve as big data scales and image resolution increases.
Figure 1. Oil tracking analytics from satellite imagery intelligence
Source: Via Satellite
Figure 2. Satellite image suggested OPEC stockpiles are 11% higher YoY
Source: Orbital Insight Inc., Bloomberg
UBS investments was one of the very first to use satellite data to predict Walmart’s quarterly earnings in 2010, which count the number of cars in Walmart parking lot each month from the image captured and cross-referencing with the unemployment numbers. Since then, satellite imageries intelligences are also used by some commodities traders and hedge fund managers for making a more sophisticated forecast on the supply, demand and risk exposures of soft and hard commodity products. Now, the scope of use even expands to traditional asset managers for sector and asset allocation, government bodies for preventing illegal mining and corporates in consumer, financial services and agriculture industries for supply chain monitoring, risk control and productivity improvement etc.
Satellite images are part of the big data revolution, and whether these data help to earn an excess return is always a question for many investors. The answers should depend on how wide spread these satellite feeds are, the cost and the consistency of these feeds. Of course, we see big data analysis still have ample room to improve. We can expect as this new technology getting more mature and the cost of data has gradually come down, more asset managers will realize the trading edge of satellite data. Investors with more information gives them a niche for making a better forecast, providing them an opportunity to invest and gain an excess return before the general public received the official data. For the time being, satellite imagery or big data may not give much direct benefit for public investors, but to ride on this trend, investors could consider investing in companies with big data technologies or have related business. If stock picking is too complicated for some investors, then an easy way is to simply investing in big data-driven funds* or even big data/technology fund as part of your portfolio for long-term investment.
Figure 3. Indxx AI Big Data Index outperformed Nasdaq since May 2016
Source: Bloomberg
*Funds that used big data analysis and machine learning to hunt out alpha. |
