Who is the most influential person in the secondaries market? According to AI-powered chatbot ChatGPT, that honour goes to Joe Baratta. 

Blackstone’s global head of private equity is undoubtedly an influential figure in alternatives, if not one of the most influential. Most would agree, though, that while his firm’s secondaries unit did raise the world’s largest secondaries fund at the beginning of this year, on $22.2 billion, Baratta himself isn’t exactly involved in the market.

ChatGPT was both close, and wildly off.

The private equity industry is grappling with how to incorporate the use of generative artificial intelligence – a large language model that uses neural networks to identify patterns within existing data to create content. According to a June survey of 110 investors conducted by Coller Capital, 60-75 percent of them think the use of AI will become significant in private equity investment processes in the next five years.

For the secondaries market – an industry worth around $106 billion last year, according to estimates by Campbell Lutyens – the notion that generative AI could be used to give buyers an edge, help make LP secondaries sales more efficient and bring secondaries market trades ever closer to the real-time nature of over-the-counter public markets securities trading is a truly enticing one.

“This is the first time in 10 years I’ve seen this level of innovation and people really thinking about [the application of AI to secondaries],” says Ross Hamilton, managing director of private equity partnerships at Partners Group.

Yann Robard, managing partner at Whitehorse Liquidity Partners, who also founded CPP Investments’ secondaries programme in 2007, has predicted we will see $1 trillion in secondaries market trades by 2030, while Vincent Gombault, who founded Ardian’s secondaries programme, expects the market to grow to $2 trillion by the end of the decade. However, multiple reasons are cited as roadblocks preventing this market’s growth.

First and foremost is the perceived lack of capital to back such transactions. According to data from Greenhill, there was around $225 billion of dry powder for secondaries at the end of last year, including near-term fundraising and leverage – enough for around 24 months of activity, based on the prior 12 months of deal volume.

Another is the dearth of people with secondaries expertise to help grow the market. Consequently, buyers, advisory firms and law firms are continuing to clamber over each other in the fight for human capital.

In conversations with more than a dozen secondaries buyers, advisers and lawyers for this report, one theme was repeatedly referred to as among the largest challenges preventing wide-scale growth: the lack of standardisation in GP reporting and LP stakes transfer processes. This disparity is making it difficult for technology to be applied and help drive greater efficiency, and therefore volume.

“If investors knew they could go through a turnkey process that included standardised legal paperwork, there would surely be more participants, which would allow for more transactions to happen,” says Chris Jeffery, general manager at Palico, an online marketplace that matches buyers and sellers of LP stakes. More participants brings more data, which in turn brings better visibility on pricing, he adds. “All this starts with having a process that sellers can rely on being streamlined and standardised.”

 

 

 

 

 

 

 

“Now we have entered the mass adoption stage of the market… An increasing pace of adoption is what’s happening right now”

Adrian Millan,
PJT Partners

Over the past 10 months, at least three firms have launched within the secondaries industry that claim to be using technology to help with LP secondaries trades – the sale of stakes in existing private markets funds. One is London-headquartered Clipway, a firm founded by ex-Ardian boss Gombault. The firm has assembled a team of former Ardian executives, as well as two data scientists who used to work at Coller Capital. One of these former Coller execs had founded RockSling Analytics, a front-office data tool that pitches itself as “data science for private markets”, according to its website.

Clipway is understood to be using machine learning, data analytics and artificial intelligence to help it model and price LP portfolios. Another firm is Bellevue, which hired a trio of investment professionals from Partners Group at the end of last year. In May, the firm partnered with Accelex, a software-as-a-service platform that specialises in alternative investment data acquisition, reporting and analytics, so it can streamline its pre-deal due diligence process.

“A lot of the players in the market right now still rely on Excel,” Lars Honegger, managing partner at Bellevue, tells Private Equity International. “That has been one of the pushes on our side – to get rid of Excel, to have a database set up where we can easily and [quickly] get information about the deal and see it immediately on a dashboard.”

The advent of the programmatic seller

What is the obvious path to growth in secondaries – a more frequently used market, or a more widely used one?

According to research by PJT Partners, shared with Private Equity International, 50 LPs have accounted for 50 percent of secondaries market deal volume over the last decade. Whereas predictions for the secondaries market’s growth have mainly focused on wider adoption across a greater range of institutional investors globally, PJT’s research suggests deeper adoption among existing sellers could be the key to the secondaries market’s growth – and technology will be the driver behind this.

“There have been repeat programmatic sellers divesting assets on a regular basis that implemented secondaries technology as core to their portfolio strategy over the last decade,” Adrian Millan, a partner at PJT Partners in New York, tells PEI. “They bring on new vintages, make changes, modify.” If first-time sellers adopt that same programmatic approach to their private markets portfolios, the multiplier effect could lead to the coveted $1 trillion of annual market activity, he adds.

Millan likens the adoption of secondaries market technology to the diffusion of innovation theory, which was developed by communications theorist Everett Rogers in 1962 to explain how a product gains momentum and spreads through a specific population or social system over time. The theory posits five ‘adopter’ categories: innovators, early adopters, early majority, late majority and laggards.

In the secondaries markets, banks and financial institutions have been the innovators, tapping the secondaries market due to forced sales. The 50 LPs driving half the market’s past decade of volume – institutions including New York State Teachers’ Retirement System and the California Public Employees’ Retirement System – are the early adopters.

“Now we have entered the mass adoption stage of the market,” says Millan. “By 2030, [we’ll be] at the top of that bell curve. An increasing pace of adoption is what’s happening right now. And in the last three years, I would argue that we’ve entered into a new phase of secondaries market activity: we’ve taken a step function from early adoption to more mass implementation.”

More effective use of technology will help with pricing in the secondaries market – something that should help shorten the cycle time of transactions. This could then enable multiple sales per quarter, GP consent notwithstanding.

“You tighten transaction cycle times, coupled with faster buyside pricing and GPs being able to process more transfers, and then that’s the virtuous circle that allows for increased volume to happen,” says Millan.  

Three-stage process

In general, LP fund stake sale processes can be divided into three sections: setup, marketing and closing. In the setup phase, buyers prepare information and analysis on funds in the event those assets become available on the secondaries market. In the marketing phase, intermediaries invite buyers to bid on portfolios of LP stakes. It is in the closing phase – especially as it relates to standardised fund stake transfer agreements – that there is the most opportunity for the application of generative AI, says Adrian Millan, a partner in PJT Partners’ capital solutions group.

“That will unlock and accelerate the market. That is the hurdle, but that is the place that is the most ready for optimisation of trades.”

When an LP wants to sell a stake in a fund, it must ask for the consent of the GP managing the fund. In some cases, the sponsor will also have restrictions as to which buyer it will allow to acquire the departing LP’s stake. The legal documents to facilitate such a process – known as transfer or subscription documents – can differ widely among GPs, meaning at least three sets of legal counsel – one for the buyer, one for the seller and one for the fund in question – must read these documents. Transfer documents can also vary in length between 60 and 100 pages.

“There’s a combination of technology that can be useful if you’ve got a lot of documents that you have to spit out that are quite similar,” says Jacqueline Eaves, a partner at law firm Goodwin. “I think that’s where you can get a lot of bang for your buck with technology solutions.”

Non-disclosure agreements – where parties agree what level of information about the underlying assets cannot be disclosed – are another stumbling block preventing secondaries trades from happening faster. Some boutique buyers are known for signing and returning NDAs without appearing to have even read them; others, such as large, listed asset managers, may take a week to respond and will often request certain clauses be removed.

“The NDA process is a real pain in the butt,” says a director at a boutique secondaries intermediary. An increasing number of brokers in secondaries trades are establishing standard NDAs to avoid having to create bespoke ones. “That’s a great approach and that’s something everybody should be doing. But at the same time, in a large, complex portfolio [sale], there’s always going to be a couple of GPs where your NDA isn’t good enough for them.”

This can result in a situation where the same portfolio sale has multiple NDAs. And when NDAs are sent out to prospective buyers, those buyers’ counsels may request changes before they’re signed, adding even more time to the process.

“Getting more of a streamlined legal standard in place for that would go a long way,” says the director.

Mike Bego, founder at lower-mid-market-focused secondaries firm Kline Hill Partners, says his firm is planning to build out its middle-office function and will use AI in transfer documentation processes, project management, number processing and communications with the seller.

“We have transferred 1,500 limited partnership interests to date – that’s over a relatively short period of time,” Bego tells PEI. “We’re doing over 110 deals a year. We have a massive volume of steps that we take to close all these transactions on a quarterly basis, and so we’re looking for advanced software tools to help us manage that process – manage the documentation handling and make it more efficient.”

“[Document generation is] where you can get a lot of bang for your buck with technology solutions”

Jacqueline Eaves,
Goodwin

This frustration around wanting to deploy technology to free up human resources was a common thread in discussions with market sources for this report. Bradley Mandel, a partner at law firm Winston & Strawn, says junior staff members are prevented from progressing as quickly as they could as they’re often tied down with menial tasks.

“A lot of this work has become so voluminous that it ties down our junior associates,” he says. Of the $7.5 billion in secondaries transactions his law firm advised on last year, 83 percent were LP portfolio sales. Junior staff aren’t left with a lot of bandwidth to get involved with purchase agreements and other key legal documents so they can understand and learn about deal processes from inception to close, Mandel adds.

“That’s definitely been a challenge for us. We certainly welcome technological developments that will ease that burden on them.”

So why hasn’t there been more progress on standardisation within this part of the secondaries process? According to market sources, there’s little incentive for GPs to change their behaviour and report information to LPs in a way that makes secondaries trades in and out of their own funds more efficient. Sponsors also report terms in different ways: a GP may use the term ‘GP commit’ in its quarterly report, while another may refer to this as ‘carry reserve’, making it difficult for software to process data in a standardised way. 

The lack of common agreement around how recallable distributions – whether the legacy LP or incoming buyer is on the hook for any distributed capital owed back to the fund post-secondaries sale – are described in sales and purchase agreements also means this part of the closing process is difficult to automate.

“It’s a very manual process,” says Chris Davies, managing partner at Bellevue. This problem becomes even more acute when it comes to withholding tax issues such as ECI – effectively connected income in the US, which causes secondaries processes to slow, he adds. “You’ve got all these complicated ECI tax withholding positions and you’ve got indirect transfer taxes. If that can be automated or semi-automated so that you can push a button and attribute a value or a risk to that, that [would save] a lot of back and forth between buyers and sellers.”

Software tools

Market participants we spoke to said they use a range of software tools to help automate menial tasks. Kira and Ontra are two examples of tools used to help with processing legal documents. Others, such as LP Analyst and Luminance, can be used to analyse the underlying data in a fund or manage a data room, while sources PEI spoke to also mentioned eFront, Burgiss, FIS, S&P Capital IQ, Cobalt and DealCloud as tools they use to help in LP secondaries processes.

Goodwin has developed its own tool that creates transfer documents, consent letters, board minutes and relevant filings. “Being able to generate those documents at the press of a button is extremely helpful. That, coupled with our use of automated subscription document software, definitely has driven efficiencies,” says partner
Andrew Pollock.

Ardian, which has built out a dedicated data science team, uses a proprietary system it developed in 2020 called iBIP – a business intelligence portal that automates cashflow projections from funds in its portfolio.

“The idea was to capitalise on all this volume of information that we gather and have the team [instead] doing their work of investment and pricing,” says Arnaud Mercier, a managing director at Ardian. The firm prices 500 funds on a quarterly basis and initially it wanted to automate some of the low value-add tasks. Now, iBIP is used to monitor positions and helps inform Ardian around decisions on whether it may want to recapitalise assets at the portfolio level or exit its positions.

At Partners Group, the firm has developed an in-house generative AI called Primera GPT, which allows it to automate text scraping from GPs’ quarterly reports. The program is “very good” at sourcing and interrogating data – a low value-add task that would otherwise be undertaken by humans, according to Hamilton. While it is still early days, Hamilton’s team has been challenging Primera GPT to see what its applications could be.

“[AI] doesn’t make the same mistake twice… so it’s like the best junior associate ever. You just train it one time and it picks up”

Ted Cardos,
Kirkland & Ellis

Specifically as it relates to modelling funds in its LP portfolio business, Partners Group is hoping the tool could help make private markets secondaries trading more similar to how information is disseminated and pricing is established in public markets. For the GPs in Partners Group’s portfolio, the task of keeping data and figures up to date is currently undertaken by relationship managers who are responsible for a given GP. Where technology can play a role is in automatically updating Partners Group’s database of funds and summarising and preparing data.

“We are trying to get as close to real-time data and pricing as possible, so that the minute you open up the model on Fund X, all of the latest news and developments that are in the firm and outside the firm are captured there,” Hamilton says. Generative AIs like Primera GPT should bring that goal closer, he adds.

Digital view

With many firms using specialised technology to attempt to make the LP portfolio market more efficient, is it inevitable that the future of the LP secondaries market is a digital one, where second-hand fund interests are traded in the ether? Not so, says Dan Nolan, former head of private markets at Swiss interdealer-broker Tradition, who joined Warana Capital this year. There’s a certain degree of salesmanship that comes with brokering secondaries trades that gives comfort to buyers and sellers that no computer can reproduce – at least, not currently. 

“I’ve worked in this industry for 12 years, and I can tell you right now that if you just shove everything into a data room and don’t have any architecture to it… it would be like selling your car without washing it – you’re not going to get the best price for it,” Nolan says.

According to another director at a boutique investment bank: “The challenge with technological innovation in some areas is that the beauty of the market is that it is such a qualitative, nuanced area where it’s really a matter of perspective. It’s almost a ‘one man’s trash is another man’s treasure’ kind of philosophy for a secondary [trade] to work – each party’s got to have conviction that the other party is a little bit wrong.”

These are elements of a transaction that can’t really be captured in an algorithm, the director says.

It is this human element of secondaries trades that many market participants say is the reason digital trading platforms for LP fund stakes haven’t ever really taken off. While increased standardisation may bring about the secondaries market’s next stage of evolution, humans will inevitably still be needed to make sense of data – especially when it comes to acquiring large portfolios.

“AI can help you go a bit faster for getting some data,” says Ardian’s Mercier, “but no one is just going to press a button and say, ‘OK, I’m going to buy $1 billion of NAV of a portfolio’.”

The future is here

At the time this report was going to press, Ardian spin-out Clipway is seeking $4 billion for its debut fund and was expected to have raised almost $1 billion as of the end of July, according to reporting by affiliate title Secondaries Investor. Bellevue was understood to be seeking up to $300 million for its debut fund, while Palico, the online marketplace, reported a 50 percent month-on-month growth in the first half of the year. It also expects to have at least $500 million in net asset value on its platform by the end of this year, according to a July statement. Tap, a digital trading platform for LP stakes worth sub-$100 million, had secured $3 million in funding from VC firms and tech executives as of June, and had $500 million in NAV for purchase on its platform in the first quarter of this year.

Asked why he thinks his platform will succeed where others haven’t in the past, Jeff Leathers, Tap’s founder and chief executive, tells PEI it’s simply a matter of timing. “No one has taken a stab at this in five-plus years, and the market has grown four times in the last five years. There’s been a lot more standardisation and there are a lot more secondaries funds, a lot more adoption among LPs who are thinking about buying as well.

“The market has matured in a way that makes it worthwhile to take a stab at [using tech in LP stakes trades]. I wouldn’t say I think there are things that those [first] guys didn’t focus on, but mostly it’s just they were too early.”

How is AI’s legalese?

When adopting AI and other technologies, law firms have their own set of issues to consider, writes Madeleine Farman 

While artificial intelligence has not been adopted formally by the law firms’ secondaries practices that Private Equity International has spoken with, there are a number of areas legal experts have identified where the technology could make a significant difference in LP-led transactions.

Anti-money laundering and know-your-customer exercises frequently came up as areas where AI could be used to streamline workflows. 

Document and ancillary work around withholding tax, LP callbacks and right of first refusal (ROFR) –sometimes seen in the venture space, where GPs may need to comply with offering its interests first to other LPs within the fund before selling to a third party – could also be disrupted, Ted Cardos, a partner at Kirkland & Ellis, tells PEI. 

There are certain legal tasks that AI is well equipped to tackle, Cardos says, with due diligence being a great example. What the technology is essentially doing is generating initial reports, which show lawyers what they are meant to be focusing on. 

“Like the ROFR example – you can train AI to say, ‘Look for transfer provisions that contain these words, these concepts, and identify whether there’s a ROFR or not’… [It’s] not going to say, ‘There’s no ROFR here’, [it’s] going to… extract from the long-form documents – ‘Here are the relevant provisions’.”

There could be some false positives, but as you train the technology, it will get better, Cardos explains. “It doesn’t make the same mistake twice… so it’s like the best junior associate ever. You just train it one time and it picks up.”

The high-level principle of automation and technology to make things more efficient is very compelling, says Ed Ford, a partner at Simpson Thacher & Bartlett. However, its use in cases where it doesn’t rely on a human check is unappealing. “We obviously need to maintain quality control. This is new to us, so it’s hard to trust it. And the clients have said exactly the same thing.” 

Efficiencies using technology are already being rolled out across law firms, with outside providers offering products and law firms creating their own platforms. Where this is being implemented, checks have been put in place before the software is rolled out for client use cases.

“There’s an entire team [that’s] tasked with doing a lot of scrubbing on the technology from an IT perspective [for] data security and the like,” says Isabel Dische, a partner at Ropes & Gray. From there, a subset of users will be tasked with pressure testing that technology.

Most firms are taking confidentiality and regulatory obligations seriously, “and appropriately so”, Dische says. She adds that the industry will take a conservative approach to the roll-out of AI, as well as technology more broadly. 

Ultimately, AI could save law firms time and, inevitably, save their clients money. 

“Some people say, ‘Oh gosh, you guys shouldn’t like this AI stuff because it takes away hours that you would otherwise be able to sell’. But the reality is nobody likes doing that work because it’s very repetitive,” Cardos says. 

With AI collecting the information that is important for lawyers to focus on, it saves individuals “digging through a 100-page document to say, ‘Hey, where is this ROFR provision?’ I’m not adding value there,” Cardos says. “If [AI] can just say, ‘It’s provision 7.6’, well, I’ve just saved myself 10 to 15 minutes, and therefore rolled forward 100 times.”