For decades, the world of finance has been preoccupied with how investments can be bundled. Mutual funds, structured products and exchange-traded funds (ETFs) are just some of the formats on offer. Now the industry has a new form of packaging.

“Technology is the new wrapper,” says Erkko Etula, the co-founder and CEO of direct indexing start-up Brooklyn Investment Group.

Etula is one of a band of financiers, quants and technologists who believe the second wave of a revolution is underway within the industry. The first wave has already resulted in a more personalised approach to finance. In the second, individualised portfolios, run largely by computer, will increasingly supplant the methods used by old-school funds.

One feature of this second wave has been the rapid growth in direct indexing, a type of separately managed account (SMA). Investors, usually wealthy individuals, concoct a bespoke portfolio using a menu of available indexes. A customer might, depending on their taste, mix an element of value investing with, say, a tilt towards sustainable companies.

Direct indexing has existed on a small scale for decades, and until recently went mostly unnoticed. However, between December 2020 and July 2022, JP Morgan, BlackRock, Vanguard, Morgan Stanley, Pershing, PGIM, First Trust and Franklin Templeton all acquired direct indexing firms in a bid to supercharge their efforts in the area. Brooklyn did a deal last December with Nuveen Asset Management to create a direct indexing portfolio with exposure to equites, municipal bonds and other asset types.

Technology is the new wrapper

Erkko Etula, Brooklyn Investment Group

The firms are responding to increasing demand from investors. Assets in direct index products reached $462 billion as of Q1 2022, according to Cerulli Associates. The research and analytics firm published its white paper on direct indexing in 2022 with Parametric, a veteran direct indexing provider that is now part of Morgan Stanley. Cerulli reckoned the figure could grow by more than 12% a year and top $825 billion by the end of 2026. It added that the value of assets held in direct indexing products could potentially surpass the value of those held in ETFs.

Now the push is towards further democratisation of the product. In early 2022, Fidelity announced that it would be offering direct indexing to customers for a $5,000 account minimum. That same year, Altruist, a platform provider for financial advisers, said it planned to offer direct indexing for only a $2,000 minimum. Parametric offers customised accounts to investors with more than $250,000 but expects to lower that to $100,000 in the near term. Towards the end of the year, it aims to offer run accounts for individuals with just $25,000 invested, though with less customisation.

The sector’s growth is being aided by technological advances and changes in market structure, and its fans say it will become the dominant form of investing in the future. But for this to happen, technologists will need to eliminate the bottlenecks in direct indexers’ workflows. This will mean accelerating their mostly automated investing processes at the points where they pass unavoidably through human hands.

Tom Lee, CEO of Parametric, says the increased customisation attracts people to direct indexing, but delivering it at scale will require a “huge technology lift”.

Avoidance is the issue

Direct indexing allows managers to algorithmically harvest tax losses – a process managers describe as “tax alpha” – while minimising portfolio tracking errors. The investors sell off assets that have fallen in value to generate tax losses; these losses can then be offset against the capital gains tax due on assets sold that have risen in value.

Delaying taxes in this way is beneficial because the investor is likely to pay lower taxes after retirement. And in the US, any outstanding capital gains tax liabilities are wiped away when someone dies.

The format also helps investors avoid paying taxes when funds are losing money. Brooklyn’s head of equities Antti Petajisto says capital gains distributions on equity mutual funds were typically 5-10% in 2022, at a time when the market was actually down 18%.

In one hypothetical scenario, Brooklyn estimates tax-loss harvesting could enable a direct indexing product to outperform a mutual fund by 225 basis points a year, and an ETF by 175bp.

A client might want something that reflects their values. Perhaps they don’t want to invest in a company that manufactures firearms or produces alcohol or tobacco

Tom Lee, Parametric

Petajisto was at BlackRock for five years, where he spent part of his time devising systemic investment strategies. Not only is tax alpha large in magnitude, he says; it is also predictable and lowly correlated with other alpha sources.

Etula says tax alpha can be conjured up within other types of funds, though much less effectively. In a mutual fund, for example, investors enter and exit at different times, and thus incur different capital gains tax liabilities.

The benefits of direct indexing go further than saving on taxes. The approach enables investors to customise portfolios with combinations of passive and thematic indexes and, potentially, active strategies. It could make it easier, for example, for investors to meet their environmental, social and governance investing goals without having to rely on blanket exclusions from a portfolio.

Lee, who started out at what is now Parametric in 1994, recalls how the firm first got into the field and how it has changed. “Our first client wanted index-like exposure with a tax overlay,” he says. “But a client might want something that reflects their values. Perhaps they don’t want to invest in a company that manufactures firearms or produces alcohol or tobacco.”

He adds that Parametric now also customises SMAs in line with clients’ retirement plans: “Maybe you work for Google and you have a lot of unvested Google stock. You don’t want a lot of technology in your index. We can customise for something like that.”

Missionary work leads to conversions

“A large part of our effort with clients over the past 10 years has been educating them on the case for direct indexing, a product they are typically entirely unfamiliar with and have little to no understanding of how it works or the value it can bring to a portfolio,” says Lee. He says this period has been one in which the product’s champions were carrying out “missionary work”. 

“More recently, as more competing firms have entered the direct indexing market, our conversations with clients and their advisers have evolved to talking about how our offering is differentiated from our peers’,” he says. “We are less frequently called on to explain what direct indexing is or the value it brings to clients.”

Changes in markets and advances in technology have accelerated the product’s progress. Most domestic trades in the US are now commission-free. Previously, brokers such as Charles Schwab would charge around $50 a trade, an amount that made it prohibitive for small or individual investors who wished to run portfolios comprising hundreds of positions. 

There are certain things you cannot automate, where you need human expertise because that’s how the client experience is enhanced

Manju Boraiah, Allspring Global Investments

“That to me is one of the biggest drivers of why direct indexing has gotten so much more popular,” says Kyle Birmingham, portfolio manager at Quorus, a direct indexing start-up that provides a platform for asset managers and financial advisers. “The market has become accessible to a broader swath of clients.”

Cloud computing has also brought down the cost of data processing, allowing for the rapid scalability of the optimisation process. Allspring Global Investments, previously Wells Fargo Asset Management, is able to run tens of thousands of account optimisations in a few minutes thanks to technological advances, says Manju Boraiah, the firm’s global head of systemic fixed income and custom SMA.

Another important part of the story has been the increased trading in fractional shares, which has made it cheaper to widen access to smaller portfolios. Historically, it was more expensive to own an entire index, but fractional share trading has enabled investors to purchase partial shares and thus bring down the price of direct indexing. 

That said, Cerulli Associates estimates that, so far, only around 5% of assets are managed through direct indexes. 

“It seems clear that increasing demand for personalisation and tax efficiency will continue driving assets from ETFs and mutual funds into tech-powered SMAs,” says Brooklyn’s Etula. “Tax-managed accounts will take market share from mutual funds and ETFs because the difference in after-tax return can be several percentage points.” Those funds today manage around half of the $8 trillion in registered investment advisers’ assets in the US market. “If you’re a fiduciary, it’s a no-brainer,” he says.

Grabbing this opportunity is the main focus right now. “It just comes down to scaling that personalised solution across the entire client base,” says Etula. “Technology did not enable that in the past, but now it can. The opportunity is some fraction of those trillions, which is also a number that keeps growing. It’s gigantic.”

People power

How that might happen may prove to be complicated. Tech alone cannot deliver the direct indexing vision, says Allspring’s Boraiah.

“There are certain things you cannot automate, where you need human expertise because that’s how the client experience is enhanced,” he says. “Ninety percent is automation. Ten percent is human expertise. It’s art and science. But the last mile has to be driven by a human. You need a combination of high touch and high tech.”

Boraiah compares his firm’s technology set-up, dubbed Remi, to an “autopilot”. He says it can manage portfolios alongside human pilots who watch out for things the system might not catch. Each day, the system checks individual retail SMAs to see whether they should be rebalanced across strategies. It then flags up to portfolio managers those accounts that need rebalancing and the likely actions that will need to be taken. 

Accounts deemed in need of attention – perhaps one out of every five – are queued for optimisation, which takes place for multiple accounts in parallel on the cloud. “The human comes in to do sanity checks before they push the button to say, ‘trade these assets’,” says Boraiah.

The obstacle to scaling SMAs, he says, is the “bottlenecks in the ecosystem”, such as onboarding new accounts, reporting to financial investment advisers and handling queries. There are 300,000 financial investment advisers operating in the US, so firms can quickly become overwhelmed by the volume of even simple communications. 

Allspring’s Remi system has completed requests for proposals (RFPs) worth close to $15 billion, by market value. Firms might hope to answer five RFPs to win a single account, Boraiah says: “For 1,000 accounts, you have to do 5,000 RFPs – that’s the scale we need to be prepared for. And you need a digital medium to do that.”

Allspring launched a financial adviser portal last year that aims to streamline the RFP process. Boraiah says the firm is exploring the use of generative artificial intelligence (AI) chatbots to deal with financial advisers’ enquiries. 

Parametric also relies on sanity checking by human eyes. “There are always fringe cases – there’s a large cash flow, the client has a value-based restriction on, but there’s a tax-loss harvesting opportunity,” says Lee. “How do you work between those objectives and understand the overriding desire of the client portfolio manager? You have to get involved.”

Automation plays a big role at Quorus too, though not fully. “The optimisation process is all heavily automated,” says Birmingham. “For every single account, all the data flows in from the custodian and is reconciled with our internal database. Then the optimisation is automatically generated for every single portfolio. At that point we’re evaluating those accounts, looking for exceptions – that may need human verification – and then the trade reconciliation process is also automated.

“We think it’s important to have a portfolio manager at the switch to handle corporate actions or even just issues at the custodian level.”

However, some in the industry think existing forms of automation fall short of what is possible. 

Etula says the number of decisions an asset manager must make to run managed accounts grows exponentially with the complexity of the mandates, meaning rules-based automation is only possible up to a point. 

We think it’s important to have a portfolio manager at the switch to handle corporate actions or even just issues at the custodian level

Kyle Birmingham, Quorus

Brooklyn’s innovation is to hand only some parts of the task to AI, thus giving more leverage to the human portfolio manager. The firm has built AI and machine-learning models that use technologies that have also been applied in large language models to do aspects of the portfolio manager’s job, though still under human supervision.

A typical human SMA portfolio manager might oversee 1,000-5,000 portfolios, and look in detail at between 50 and 100 each day. Brooklyn believes its system can enable a portfolio manager to manage 10 times that number.

“A rules-based system works great if you can come up with rules for every specific situation,” says Etula. “But when you deal with a multitude of different objectives, different benchmarks and different sets of preferences – one client wants to do a tax-neutral sell-down of their concentrated position, say – you need artificial intelligence to replicate part of the human thought process, to ask: ‘For this fact pattern, for this account on this day, are there any urgent actions? How should this portfolio be treated?’

“You have tasks where, frankly, a human cannot process that much data and keep track of so many things.”

Quorus is also looking at using AI to help determine which accounts to review and which ones need portfolio manager approval. “We see it as part of our operational effort, not necessarily part of the investment decision-making process,” says Birmingham. 

Old hands versus young guns

Who will win the race to solve these problems is hard to say. 

Start-ups are arguably more nimble and have the opportunity to build their tech stack from scratch and focus on making it scalable. Brooklyn has 13 software engineers on its staff of 21. 

Incumbents, though, have more heft. Parametric’s Lee says: “I am happy that we have the experience, knowledge, the human capital that we have, and the client relationships that we have.” 

He reckons the field will narrow to four or maybe six players as firms baulk at the up-front costs of direct indexing and perhaps become frustrated by slower growth than they had been hoping for.

Finance is about trust as well as tech, and incumbents own most of the customer relationships right now. 

That said, Lee says he “wakes up every day afraid of the competition”.

“Being the largest player can be somewhat of a disadvantage, because we’re driving an aircraft carrier and other people can be driving a boat,” he says. “They can just make a quick change where we may have to consider our larger book of accounts.”

Parametric is currently migrating its platform to the cloud – a process that it says takes time. “You can’t make mistakes,” Lee says. “Functions will migrate over time, and we will have to run some things in parallel for a time before we unplug legacy systems.”

Birmingham says Quorus has worked hard to build a backtesting tool to show financial advisers how its optimisation software would have added value to various strategies using historical data. 

“The biggest hurdle we see when we go out and talk to advisers is getting them to understand the value to them and their clients,” he says. 

Brooklyn’s Etula acknowledges the challenge. “The one thing that’s perhaps impossible to automate is the ability of a skilled investment adviser to build trust,” he says. “That’s why the investment adviser is always in the driver’s seat.” His firm is testing its models with a view to releasing empirical evidence of their reliability in the coming months. He says Brooklyn employs “guardrails”, using AI agents that work within its systematic portfolio management process to make sure its AI does not miss portfolios it should be flagging. 

Perhaps for that reason, newer entrants to the space are as keen to work with incumbents as to challenge them. 

Brooklyn wants to work with other asset managers and allow them to white label its technology, as it has done with Nuveen. “Nobody needs to know who Brooklyn is,” Etula says. “We are here to power the brands of our asset manager and investment adviser clients.”

Quorus sees relationship-building as a key element as well. “On the asset manager side, it’s a web of competition,” Birmingham says. “But our plan isn’t to have proprietary asset management. We’d like to be a partner in this space, not a competitor. And I think that’s a compelling offer for some of these asset managers.”

Editing by Rob Mannix and Daniel Blackburn