The AWS of Finance
Plus: Russian Fintech, Fintech and Big Tech Credit, Abraham Wald and the FinCEN Files
|Marc Rubinstein||Sep 25|| 15||2|
Welcome to another issue of Net Interest, my newsletter on financial sector themes. Every Friday I distil 25 years experience of looking at financial institutions into an email that explores key themes trending in the industry. I’m grateful to all of you who have signed up, but don’t keep it to yourselves! Tell your co-workers, friends, family, neighbours. And do get in touch by hitting reply.
The AWS of Finance
We all love a good origin story. Steve Jobs assembling kit in his parents’ garage; Mark Zuckerberg coding in his dorm room; Reed Hastings returning his overdue video rental to Blockbuster. These stories help animate the companies they underpin.
In finance, such stories are, sadly, scarce.
One exception is Blackrock.
In 1986, Larry Fink led the mortgage department at First Boston, a large investment bank, now part of Credit Suisse. He’d joined as a graduate trainee and worked his way up to become tipped as an eventual CEO. But he made a bet that interest rates would rise and they didn’t—they fell, losing the firm $100 million. Fink was pushed out and went on to create Blackrock, today the largest asset management firm in the world.
We’ve talked about Blackrock on the periphery of Net Interest in the past—about its size ($7.32 trillion in assets) and about its dominant position in the field of passive asset management. We haven’t yet gone deep. Which is remiss, because in the narrative we’ve established about the plumbing of markets, Blackrock is a key player. It stands alongside the exchange groups as a major beneficiary of the shift in power away from investment banks.
Blackrock’s core business is investment management. Of the $15 billion or so a year the firm makes in revenue, around $12.6 billion comes from investment management. Investment management is a fairly simple business to understand. Blackrock manages money on behalf of others – pension plans, endowments, sovereign wealth funds, central banks, corporations, financial institutions, regular people – for which it takes a small cut, averaging around 0.17% of the assets managed. That cut’s been going down over recent years as a result of competition and a bit of a mix shift, but as long as total assets under management go up, revenues will rise.
In the investment management business, Blackrock has the benefit of scale. Few other firms can take as low a cut as 0.17% and keep the lights on. Blackrock is able to take that 0.17%, absorb a $10 billion cost base and keep a 38% operating margin for itself. Unless its products go out of favour, its scale sustains a competitive advantage. And its products are not likely to go out of favour because there are so many different types of them (including lots of passive tracking products which are very much in favour).
But Blackrock is more than just an investment management firm. In its filings, the company describes what it does thus:
“BlackRock provides a broad range of investment and technology services to institutional and retail clients worldwide.”
Investment services we’ve discussed, but what’s all this about technology services?
Look closely, and you’ll see another revenue stream at Blackrock, tabbed technology services. It’s smaller – around $1 billion – but it’s growing fast. Its roots lie in the origin story of Blackrock itself.
Larry Fink is no dumb guy. He didn’t just bet blindly that rates would go up when he was at First Boston. He hedged his bet, so that even if rates went down instead of up, he would be protected. Or at least he thought he would be protected. Stephen Schwarzman explains:
“But a guy from the back office who ran Larry’s computer models had made a mistake, and the hedges were wrong. Larry had made his calculations based on the wrong numbers. In a single quarter, his department lost $100 million. It wasn’t his fault; he didn’t control the back office.”
The episode taught Fink that in future he would need to control the back office.
At Blackrock he developed a risk management system that would be fully integrated with the investment process. He called it Aladdin, short for the less snazzy “Asset, Liability, Debt and Derivative Investment Network”. Aladdin’s job was to provide a comprehensive risk overview of the firm’s portfolios. It became the central system for the firm’s position-keeping, record-keeping and risk control.
By 1994 Fink’s risk management system was helping him manage around $20 billion of assets. But then the bond market sold off after the Fed began raising interest rates more than the market was expecting. Investors in bond funds, including Fink’s, suffered losses so swift and severe that the period became known as “the bond market massacre”. Blackrock’s funds recovered; by the end of the year its Strategic Global Bond Fund was down only 4.0%.
Others weren’t so fortunate. According to Rob Goldstein, COO of Blackrock, “People had a lot of mortgage securities that they had bought; that they didn't really have technology to understand what they had bought. We started getting calls from people saying, can you take a look at my portfolio and tell me what you think of it.”
On Halloween 1994 they got one such call from General Electric which was trying to sell its brokerage subsidiary, Kidder Peabody, but was unsure how to value its assets.
“We were handed a disc. It was a spreadsheet that had like 1000 rows. And it was called Michelle’s spreadsheet… Michelle was someone who worked at Kidder Peabody and as people would do trades… they would sort of shout at Michelle and she would update her spreadsheet.”
The episode made Blackrock realise that it could offer Aladdin’s services out to third parties. On completing the General Electric job (after three all-nighters in a row), the firm went on to license the platform to Freddie Mac and others. By the end of 1998, Blackrock was providing risk analytics services to ten clients, covering more than $400 billion of assets.
As Blackrock grew, Aladdin grew. The firm bought Merrill Lynch Investment Managers in 2006, taking it deeper into Europe and into equities markets. In 2009 it bought Barclays Global Investors, taking it into exchange traded funds. Billions of assets under management became trillions of assets under management.
But the firm’s big break came in 2008, during the financial crisis. Around the world, governments struggled to evaluate the risk exposure in the portfolios of their banks. The banks themselves couldn’t help due to their obvious conflicts, so governments called on Blackrock to value exposures. Aladdin was used by the Fed when it took on the assets of Bear Stearns, and then by the Treasury in the broader bailout. By the end of 2008, Aladdin’s services were used by 135 clients, covering $7 trillion of assets.
Since then, Blackrock’s technology platform has kept on growing. It has been hired for short-term projects of the kind it did for General Electric and the Fed. During the European debt crisis it was hired by the central banks of Ireland, Greece and the ECB.
More significantly, it is also used as an investment and risk management system for institutional investors globally. The platform monitors 2,000+ risk factors each day – from interest rates to currencies – and performs 5,000 portfolio stress tests and 180 million option-adjusted calculations each week. This enables it to provide an enterprise view of a firm’s risk. Investment managers who want to see what their overall exposure is to airlines or hospitality, say, can do so in a click. Alongside other features such as portfolio management, trading and operations, Aladdin positions itself as “an operating system for investment professionals.”
Today, Aladdin has over 250 clients. It has taken on 18 even since the pandemic began. Clients include asset owners like the California State Teachers’ Retirement System and other asset managers like Schroders and Vanguard.
Like many software systems, the platform is quite sticky. Implementation can take 12-18 months, but once the client's on board, they’re on board—renewal rates are in the high 90’s of percent. Fees are based on criteria such as the value of positions, number of users or accomplishment of specific deliverables. JPMorgan became a client in 2016, paying $1.5 million for integration, plus annual fees that grew to $5.3 million last year.
Aladdin’s growth comes from more clients, greater penetration of those clients, and more products. Last year Blackrock spent $1.3 billion to acquire eFront to inject illiquid investment expertise into the platform. eFront is a leading software and solutions provider in the alternative investment management space. The firm also rolled out Aladdin Wealth to service the wealth management market. It currently has 16 clients there, including Morgan Stanley.
Too Big to Fail
Aladdin started out as a cost centre deep inside Blackrock (its “central nervous system”) but has evolved into a robust profit centre. The company doesn’t break down Aladdin’s full financials (despite its size, Blackrock still operates as a single business segment) but as a software business its margin is likely very high. It’s a strategy other financial services companies are looking to pursue. Goldman Sachs has long marketed itself as a technology company, with 25% of its workforce being engineers (“more engineers than Facebook”) yet for now its financials don’t reflect one.
Blackrock faces two problems as its platform business grows. First is the conflicts it may trigger. Its investment managers have stakes in companies which are also Aladdin clients. Investment managers are familiar with conflicts of interest—their clients want investment performance; they want business performance. But this type of conflict – between client segments – is entirely different. It’s the type of conflict that investment banks normally grapple with, not investment management firms.
Second, like all platforms, there’s a point at which it gets so big it becomes systemic. Some may argue it’s already there. The company stopped reporting how many trillions of assets sit on the Aladdin platform at the beginning of 2017. The last reported number was $20 trillion. Earlier this year, the FT calculated that a third of Aladdin’s clients have $21.6 trillion on the platform. According to the FT, that figure alone is equivalent to 10% of global stocks and bonds.
The concern is that with so much money managed off the same risk system, portfolios will begin to look alike. The Los Angeles County Employees Retirement Association tendered for a risk management solution earlier this year. Their concern with Aladdin was “potential groupthink” on the basis that Blackrock’s asset management teams use the identical platform in their investment process.
Not a month after the retirement association made their decision, Blackrock rolled out a coronavirus stress-test scenario to Aladdin clients. It would be interesting to see how those clients performed in the ensuing market turmoil in late February/March and to what extent Aladdin influenced herding. Market correlations rose to 20 year highs in the period, which isn’t a surprise – they always rise in periods of stress – but what impact Aladdin may have had is a question few have asked.
The flipside is the benefit that comes from bringing together so much data on one platform. Blackrock markets Aladdin’s collective intelligence on the basis that it “gets better with every new user, and every new asset that joins the platform”. This is true of most platforms built around data. In this case clients have to trade off the value of that collective intelligence with the potential groupthink that comes with it.
One of Aladdin’s clients, European insurer NN Group, thinks it’s found a solution. In June, they said:
“Now we have a state-of-the-art platform in BlackRock Aladdin… Man and machine as a combination is an important driver for us to look at ways to generate more alpha, more investment returns for our clients.”
Man and machine. The genie and the lamp.
More Net Interest
Yandex, Russia’s answer to Google, has announced it’s acquiring online bank, Tinkoff. Yandex offers a broad range of app-based services across search, e-commerce, ride hailing, video streaming, food delivery and more. Up until June, it had a relationship with Russia’s largest bank, Sberbank, but that was unwound with Yandex buying Sberbank out of their e-commerce joint venture (Yandex.Market) and Sberbank buying Yandex out of fintech (Yandex.Money).
Since then Yandex has been looking to fill the fintech hole in its offering, which management has previously described as the ‘glue’ of an ecosystem. Tinkoff does the job, giving Yandex a credit card product (number 2 in the market with a 13% share), a brokerage product (2 million accounts) and an SME product (565,000 clients). It comes with over 11 million customers, of which 6.1 million were monthly average users in the second quarter of the year.
The model aligns Yandex with Asian internet platforms who want to do everything. Ant Group started with fintech and is now branching into other areas; Yandex is doing it the other way round. They are aligned in their view that fintech is the glue, as the piece of the platform that provides both data and liquidity. Other internet platforms have been reluctant to move quite as aggressively into fintech because of the regulatory oversight that accompanies it. But if the model takes hold, challenger banks elsewhere in the world may find themselves targets.
Fintech and Big Tech Credit
Despite all the chat about fintech credit, the market globally remains quite small. The BIS published a new database this week that tracks fintech and big tech lending volumes across the world. They estimate that the stock of fintech and big tech credit reached $303 billion globally at the end of 2018, or about 0.3% of the stock of overall credit to the private sector.
However, a lot of fintech credit is quite short-term, so the flows are higher than that. According to BIS estimates, fintech and big tech credit volumes hit $795 billion in 2019. What’s interesting is that big tech companies are the primary source of this lending, rather than specialist fintech companies. Of the $795 billion, big tech accounts for $572 billion of credit flows.
The biggest market is China. Most of the big tech lending globally – $516 billion – is done by one of the big four Chinese tech companies—Ant Group, Tencent, Baidu and JD.com. By contrast, specialist fintech platforms are shrinking in China. From a peak of 3,600 fintech credit platforms in November 2015, only 343 were still in operation in December 2019. Their new lending was only at 17% of its July 2017 peak level.
US big techs haven’t done much lending yet. The BIS mentions that Amazon did $1 billion in the US in 2018 and that Apple Card had $7 billion of balances at the end of 2019. However, even that small slice gives Apple Card a 13% share of overall fintech balances at year end, which isn’t a bad start.
Abraham Wald and the FinCEN Files
Banks were hit this week by severe money laundering allegations following the release of the FinCEN files, involving around $2 trillion of suspicious transactions. While it doesn’t look good for them, it should be noted that these files were generated by banks themselves and sent to regulators.
Looking at the data on the International Consortium of Investigative Journalists website, I am reminded of the story of Abraham Wald, a mathematician who worked at Columbia University during the Second World War. Wald was tasked with figuring out how to better protect aircraft from damage during flying missions. The initial approach to the problem was to look at the planes coming back, see where they were hit the worst, and reinforce in that area. However, Wald realised that it was the planes that didn’t make it back that harboured the most useful information.
Estimates put global money laundering volumes at between $800 billion and $2 trillion per year, equivalent to 2-5% of global GDP. Banks have a duty to report suspicious activity which is what the FinCEN files represent. Yet according to Europol, in Europe at least, reports account for only 0.1% of GDP. Either most of the world’s money laundering is going on outside Europe (yeah, right), or it’s not being detected. Even when reports are raised, there may be a fairly sizable false positive rate. Europol calculates that only 10% of reports are escalated (and barely 1% of criminal proceeds are ultimately confiscated).
Some juicy stories will emerge from this data dump. But the real story is in the planes that aren’t coming back.