Meet Aladdin, the computer “more powerful than traditional politics”
How a single computer system came to influence the value of vast swathes of the world’s financial assets.
By Will Dunn
The winter of early 1993 was a cold one in New York; the beginning of February saw the mercury drop below -13°C. Each morning of that winter Jody Kochansky would arrive at 6.30am at the Manhattan offices of BlackRock, and begin going through the printouts.
“I’m not a big morning person,” Kochansky admits. To compound the early start, the first job of the day was also the most arduous: “to take the risk reports, to flip the pages and literally to compare the portfolio as it looked today versus the portfolio as it looked the previous day, by hand.” The first web browser would not be created until later that year; “the delivery mechanism was paper”, but Kochansky and his team found a solution. “We said, let’s take this data, and rather than print it out, let’s sort it into a database, and have the computer compare the report today versus the report yesterday, across every position.”
From a simple time-saving system designed while most of New York was still in bed, BlackRock’s computer system has grown into the “operating system” for a company that has itself grown into the world’s largest manager of financial assets. The system, now known as Aladdin, inhabits multiple datacentres – warehouses filled with servers – and is used by around 13,000 BlackRock employees and thousands more people at the company’s clients, who pay for the analysis the system provides.
This is how much money Aladdin manages: if you took every last cent out of every bank in every country in the world, emptied the wallets and pockets and penny jars of all 7.6 billion people, if you rummaged down the back of every sofa and emptied every till and safe until you collected every scrap of currency in the world, you would have a pile of cash worth around five trillion dollars. The total value of assets under management by BlackRock is $6.3 trillion. But Aladdin also delivers risk analysis on the assets managed by its clients, which are valued at more than double that amount. Overall, Aladdin has an effect on the management of around ten per cent of the world’s financial assets, or around $20 trillion. Over 25 years, it has grown into a system that is directly or indirectly responsible for more than four times the value of all the money in the world.
The fact that Aladdin demonstrates about financial technology is that it is not the technology itself that creates success, but how it is used. Kochansky says broker-dealers were using powerful mainframe computers to understand the risks that applied to different investments as far back as the 1980s. BlackRock’s founders invested in new “workstations” – “cheaper computers, that cost tens of thousands rather than millions”. Their innovation was to use the cheaper computing power not only to sell securities, as others did, but to know the true value of what they were buying.
How does a computer know how risky something is? Kochansky says that the mathematics “can be fairly complex”, but that Aladdin uses “Monte Carlo simulations”, among other models, “to try to see what happens to the security under different kinds of environments”. A Monte Carlo simulation is a type of algorithm that simulates the messy unpredictability of the real world within the deterministic order of mathematics. To do this, it uses random numbers to calculate not exactly what will happen, but what is likely to happen.
When BlackRock began applying this type of mathematics to building portfolios, they were run on a single Unix workstation that was “literally the desktop computer for the trader,” Kochansky remembers. “But then at night we would use the compute resource to run our bond analytics, and the next morning the portfolio manager would get a fancy report that other buy-side organisations couldn’t get.” The “fancy report” would contain risk analytics on “everything they owned. Every single day, our portfolio managers could see the risk on their entire portfolio.” Today, when analytics are applied to everything from training athletes to selling deodorant, this would be expected, but in the early 1990s BlackRock was the first and only company to use data in this way.
“Nowadays the hot topic is fintech,” say Kochansky. “I like to think that we were one of the earliest fintechs.”
But BlackRock soon realised that the system should not just calculate risk. The next step was to use the system for “position-keeping, record-keeping, and control”. The reason for this was that at the time, with risk calculated separately from positions, analytics were always one step behind trading. “We realised that if you know how risky a security is, but you don’t know how much you own, you don’t really know your risk,” he explains. “You ultimately have to marry the risk calculations at the security level with the portfolio holdings to truly get a view of risk. We needed to know how much we owned.”
This insight was particularly important in the autumn of 1994, in the wake of a crisis in the bond market remembered by Fortune magazine as “the Great Bond Massacre”. As bond prices fell, General Electric began looking for a buyer for a Wall Street broker-dealer that it owned. “They went to other broker-dealers,” Kochansky remembers, “and said, hey, we’ve got this complex portfolio, can you provide a bid for it – in other words, how much would you pay me for it?” The bids GE received were low – “much lower than they had expected.” Kochansky says this “makes sense, because they were trying to move a big block of securities” – companies rarely get top dollar for their stock during a closing-down sale. The answer was to have BlackRock manage the portfolio, gradually selling it over time.
Kochansky and his team “literally pulled three all-nighters in a row. We modelled the entire portfolio. It was very painful, lots of coffee. But when we were done – wow.” Incorporating a broker-dealer system into Aladdin was, says Kochansky, “the moment when we realised that the platform was capable of doing a lot of things that we hadn’t even contemplated at that time.”
By the mid-1990s, the system was able to “rebalance” portfolios “in more automated ways”. Once a portfolio manager knew how risky their portfolio was and how much they owned of every security, they would look to invest in other securities, and Aladdin was designed to adjust the portfolio, to balance automatically the risk being introduced. It was in this way that Aladdin was given, says Kochansky, “the ability to interact with the marketplace. In those early days, in the mid to late 90s, the vast majority of trading in fixed-income was really based on the telephone. People would pick up the phone and offer two-year notes and that kind of thing. So as the electronification of the marketplace was happening, we were building out the market-facing technologies.”
In 1999, BlackRock went public at $14 a share. Kochansky points to other moments in the firm’s history when the company and its technological core had to be recalibrated. In 2006, BlackRock acquired Merrill-Lynch investment managers and “we suddenly became very international, and we became very equity”. But it was the financial crisis of 2008 that really proved Aladdin as a significant influence on the global economy. As the government struggled to make sense of the febrile financial markets, asking an investment bank to help value the rest of Wall Street would have been, Kochansky puts it diplomatically, “a potential conflict”. “And so I think as these governments – it wasn’t just the US government – looked around and asked who had the capabilities to gain the insight into what’s happening in these portfolios, that’s a relatively short list.” Unlike an investment bank, BlackRock does not trade its own capital. This fact, coupled with its analytical prowess, gave it an unparalleled ability to value the twelve-figure refinancing deals needed to keep the US economy afloat. In 2010, Vanity Fair reported that “BlackRock has effectively become the leading manager of Washington’s bailout of Wall Street”. BlackRock’s share price at time of writing was $521.
The simulated economy
While the world has changed dramatically since Kochansky started at BlackRock in 1992, he says that “a lot of the mindset that was created then propagates today. The principles are the same. What we nowadays call Aladdin – at the time, it didn’t really have a name – is the operating system for BlackRock.”
In 1994, when Kochansky and his team worked for 72 hours to rewrite Aladdin for the first time, “the entire effort around Aladdin was maybe 40 or 50 people. Nowadays, within the Aladdin product group we have about 1,500 to 2,000 people that contribute to Aladdin in different forms.”
Physically, Aladdin occupies three datacentres in the United States. Unusually, these are owned by the company itself rather than a third party. The company is looking at opening a pair of datacentres in Europe. Kochansky says “it’s hard to characterise it exactly by number of computers, but one way to think about it is that we are running the risk analytics on tens of millions of securities.” Every individual security is valued through “thousands and thousands of Monte Carlo simulations, and each simulation is a matter of creating an economic scenario that’s based in statistical grounding. One security manifests as millions of scenarios. So, we are running billions and billions of scenarios every single night and throughout the day.”
In a typical day, Kochansky adds, the cloud-based system will pass “tens of billions of messages” to distribute analytics to users and clients.
Every security that is analysed goes through “the economy process”, which Kochansky describes as valuation in the context of every data point BlackRock can gather than might affect the economy – “a real-time snapshot of all sorts of market information. So, anything you could imagine that’s going on in the marketplace – what are the Treasury rates, what’s the shape of the yield curve, what’s going on in equities, what’s the state of volatility. In the case of mortgages, you want to know stuff about the rate of inflation, home price appreciation trends by zip code. We call that ‘the economy’.”
Every single one of the tens of millions of securities Aladdin analyses is valued against “hundreds of thousands of data points each day”, in thousands of different ways.
Too big to fail?
In December 2013 Stanley Pignall, the Economist’s finance correspondent, said of BlackRock that “it’s unprecedented to have a single firm that has such a grip on the way that not only itself, but its rivals, look at the world.” While Pignall said BlackRock was not “too big to fail” in the sense that some banks had been – as an asset manager it is not exposed to the same risks – he said the sheer size and power of the Aladdin platform had some economists spooked.
The reason, said Pignall, was that so much value is now managed using Aladdin that if, for example, JP Morgan, Deutsche bank and some of the world’s biggest sovereign wealth finds all use the same model, “the risk is that they start finding certain types of assets attractive, or unattractive, at the same time. You’d get a herding of investors,” Pignall suggested, similar to that “in 2008, when too many people started listening to the credit rating agencies, and started buying products that were linked to subprime real estate in the US.”
Kochansky answers that “it’s important to recognise that Aladdin itself does not predict the future at all. Aladdin tells you what you have in your portfolio. It doesn’t tell you what to buy.” There are, for Aladdin, no good or bad bets – there is only risk, in varying degrees.
“The enterprise clients have a separate instance of Aladdin,” he adds, “meaning that it’s all their own data, running on separate computers, separate databases. They dial the models based on their views of the markets, their views of risk. Aladdin doesn’t dictate to them how to run their business.”
Nor is Aladdin the only system in play. “Keep in mind that the models in Aladdin are based on historical data. There are many providers in the marketplace that provide risk models based on historic data. Our models are world-class, but we’re not the only model provider in the world,” says Kochansky. Elsewhere, behemoths such as MSCI Barra and Bloomberg offer their models and systems that may influence still more value even than Aladdin. While Kochansky admits, memorably, that “$20 trillion sounds like a lot of money,”there may be, “in the grand scheme of the capital markets”, models that have still greater influence.
The idea that computer systems and financial models of this sort quietly underpin the value of almost everything in the world (everything that can be invested in, anyway) could be either frightening or reassuring. The filmmaker Adam Curtis has described Aladdin as “a kind of power never seen before… more powerful in some respects than traditional politics.”
On the other hand, if the principal aim of such a vastly influential system is the management of risk, it could be the steady hand that the markets of the future will need . A powerful stabilising technology such as Aladdin could yet be the source of “Great Moderation” that neoliberalism tried to deliver.
Either way, Aladdin is set to become more powerful still. BlackRock’s $109bn research unit, the Systematic Active Equity division, has been investigating artificial intelligence for some time, and the company recently announced that it is building a new laboratory in California to develop this technology even further. As to what kind of future this creates for the world and its economy, Aladdin itself will probably be the first to know.