Space, Money and Baseball
Even if you aren't a die-hard baseball fan, you've probably heard of Billy Beane. Depicted in Aaron Sorkin's Moneyball, Beane's story is one of the great underdog stories in modern sports history. Operating with a fraction of the budget of teams like the Yankees, Beane built one of the most cost-effective teams in baseball, guiding the Oakland A's to become the first team in 100+ years of American League Baseball to win 20 consecutive games.
How? Beane used innovative statistical approaches and went against the grain of traditional grey-haired managers. However, Beane's story isn't just a baseball story. It's a story of shifting paradigms. Beane not only proved that underdogs could overcome their limitations but showed the whole world that pockets of value exist all around us and can often be found by looking where others aren't.
Beane's story is now being told in the financial world with the explosion of alternative data. Like baseball, the investment game is an extremely competitive sport. Making only a slightly better risk-adjusted return than the market can cause massive investment inflows. Because funds' revenue comes from fees on the amount of money they manage, everyone is trying to find that edge over others to beat the market and keep their firm alive.
Believe it or not, some have found their edge by looking at earth from the heavens...
Case in point, RS Metrics
Satellite imagery is now a hot commodity for hedge funds and investment banks. The data provides such an informational advantage over the competition, its price can reach hundreds of thousands of dollars. Unsurprisingly, there are many companies solely dedicated to the collection and distribution of these images.
One such firm, RS Metrics, was founded in 2009 by two brothers, Tom and Alex Diamond. Inspired by Walmart founder Sam Walton, who used to count the number of cars in a store's carpark in order to judge its performance, Tom wondered if you could apply that method to all Walmart carparks.
He convinced Alex to leave his job at satellite firm DigitalGlobe, and take with him three years worth of satellite imagery of outdoor retailer carparks. They found the number of cars parked in the parking lots accurately predicted the next quarterly earnings release for the company. For anyone familiar with the stock market, you know the Diamond brothers stumbled across a gold mine. Their own little Peter Brand (Jonah Hill).
That next year they started their firm, RS Metrics, which now takes photos of not only retailer carparks but a whole bunch of other cool things as well, which they sell to top-class hedge funds.
One of their big products these days is called MetalSignals. It's a whole bunch of satellite images of commodity smelters and outdoor storage facilities around the world. Hedge funds can buy these photos to have a good idea of where commodity supply is and therefore how prices will move.
Money... I smell money!
The main way hedge funds use satellite data is to predict surprise earnings announcements. Let's say Elucidate hedge fund bought access to current satellite images of the Tesla gigafactories.
Looking at the number of cars exiting the warehouses, we can track the supply of Tesla's worldwide fleet. If our satellite data shows Tesla produced 1 million cars in the past year, but the market expects them to have only sold 900 000, the annual earnings release will most likely be unexpectedly positive.
So, Elucidate hedge fund will buy Tesla stock and hopefully make a sweet profit when the earnings figures are released.
But, it isn't limited to this sort of activity. Hedge funds use the data for far more short term, opportunistic trades as well.
In 2018 a train carrying 270 carriages of iron ore derailed in the Pilbara region of Western Australia. The sudden fall in supply caused world iron ore prices to surge. According to The Atlantic, a bunch of traders had a look at satellite images of the derailment and noticed the land around it was extremely flat.
They figured that getting the iron ore back onto the carriages would be quick and that hardly any cargo would be lost in the process. So, they bet against the iron ore price. When the news then broke that the cargo was back on board, their bet paid off, prices fell and they made a fortune.
Similarly, hedge funds are buying up satellite images of airports around the world. They use this data to track company private jets in order to predict mergers and acquisitions and trade on the information.
Think of it this way. Hedge funds having satellite data is like Marty McFly travelling forward to 2015 to bet on the Cubs winning the World Series. In a way, they know something is going to happen before it is announced, and more importantly, before everyone else knows.
What does the research tell us?
A paper out of UC Berkeley, On the Capital Market Consequences of Alternative Data: Evidence from Outer Space, borrowed and looked at the carpark satellite data RS Metrics collected between 2011 and 2017. The data contained images of the carparks of 44 US retailers, Walmart being one. Note, the data they looked at had been sold to hedge funds and was presumably traded upon in that period.
The paper found:
Trading on this satellite data was very profitable. If there are an abnormally large number of cars in the carpark of a retailer, we can assume revenue and therefore earnings will be above expectations, pushing the stock price up. The paper finds that buying shares in the companies with the largest increases in carpark fill rates reaps a 1.7% return above the market. Alternatively, short-selling the companies with the largest decreases reaps a 3.1% return above the market. Satellite data provides that edge we're looking for.
Hedge funds are trading on the data. Specifically, they are short-selling. The paper finds short-selling surges on stocks with abnormal decreases in carpark fill rates just before earnings announcements. Importantly, there is no surge for those companies with abnormal increases in carpark fill rates. Traders are making the right bets.
These trades are not making the market any more efficient. If hedge funds are using satellite data to make trades before earnings announcements, you would expect the stock to move toward the fair price gradually, rather than all on the day of the announcement. However, they find trading volume only increases on the day of the announcement. In other words, the fact that hedge funds are making these trades isn't leading to stocks being more accurately priced. Hedge funds are taking all the winnings, as they are the only investors who can afford the data. Individual investors and less wealthy funds can only dream of the stars.
Uninformed individual investors are missing the party. The paper shows individual investors (who don't have access to satellite data on the 44 companies) tend to buy the stocks they should be selling. These are the stocks the satellite data indicates will have surprisingly poor earnings in the future.
The research tells us one main thing: if you're a fund manager and can afford it, you really want to buy this kind of data.
A brief discussion
What if we told you satellite data is just the tip of the iceberg? In reality, we have seen an explosion in the number of alternative data sources in the last 10 years as funds seek to innovate and find their edge.
Hedge funds and investment banks don't just look at satellite imagery. They look at weather patterns, track credit card payments and even develop Natural Language Processing models to analyse the sentiment of Tweets and Reddit posts pertaining to certain stocks.
One firm, Thinknum, now tracks stock mentions on WallStreetBets, the subreddit responsible for the GameStop saga back in early 2021. Had they been collecting this data back then, huge hedge fund losses could have been avoided.
These days, fund managers are using machine learning to interpret vast swathes of alternative datasets simultaneously. Data and data scientists are the hot commodity on Wall St.
All in all, financial markets' adoption of alternative data is a good thing. If investors have a better idea of the true value of stocks, market pricing will follow suit. The more information given to a market, the more efficient it will be.
But, as we saw in the paper, we aren't seeing this efficiency quite yet. Satellite data is incredibly hard to get a hold of; it is extremely expensive and necessitates the hiring of highly skilled data scientists, a scarce commodity. As a result, only a small subset of ultra-rich, sophisticated money managers can afford it.
Alternative data will get cheaper, and with that will come greater efficiency. This will take time.
For the moment, the spoils are highly concentrated. A co-author of the paper even likened the use of alternative data to insider trading. They argue this because they don't see alternative data as public information, but as selective and expensive information to which the broader market doesn't have access. Insider trading laws seek to limit the market power and capability of individuals and firms with non-public information.
We think this is an interesting but ultimately unconvincing argument. Regulators should be encouraging sophisticated investors to incorporate better information into their investment decisions. The use of alternative data should by no means be illegal. But maybe it should be more accessible.
Moneyball
In the same way Billy Beane changed baseball, alternative data is changing investing. Scouts were replaced with computer-based player profiling. Cash flow assumptions and subjective revenue forecasts are being replaced by cold data and machine learning models.
But unfortunately, we all know the movie's ending. The Oakland A's bow out in the American League Division Series; the LA Angels win the World Series. Every team starts to use Billy Beane's method.
It's the same in the financial world. The innovators who used alternative data before everyone else have done well. But the big guns have caught up. Soon it'll be a part of almost every fund manager's decision process.
At least those who were first will go to bed knowing they changed the game...