The Promise and Peril of Tokenized Funds
Tokenized VC funds have attracted a lot of attention and a little bit of money. This article will explain what the excitement is all about, and what problems remain to be solved. It will dive into the details that professionals will need to design these funds or invest in them. I will also comment on some of the players — naming names and kicking butt.
Tokenized funds can expand VC investing beyond a small club of tycoons and make it accessible to an expansive global market.
To understand the interest in tokenized funds, we have to consider the way that VC funds are organized now. They are funded by “limited partners” (LPs) who are large investors that sign up for a ten year deal. The LPs commit to deliver a certain amount of capital. During the first years, the VC will scout for deals, and when they need money to invest, they make a “capital call” to get it. Later, when the investment is sold, they return money to the LPs, hopefully with a profit. The world of LPs truly is “limited”. There are a limited number of investors that can reliably deliver on capital calls, and then wait ten years to get their money back. VCs get burned out trying to pitch the small group of usual suspects. And, normal people can’t make these investments.
A tokenized fund would be delivered to investors as a security that can be resold as soon as a year later. This can bring in new investors. There are a limited number of investors that can commit to a 10-year deal. A greater number could invest if they had the option to sell their stakes after a shorter time period. And, we can use the global blockchain to expand into a global market of investors. And, we can design “programmable securities” that allow us to reach new groups of investors by applying fine-grained rules about who is allowed to buy the security.
In theory, a new tokenized/securitized fund structure should be optimized for investor returns, so that it can compete with the old LP structure. Then, it would add the advantage of packaging that can be resold in a secondary market.
The current generation of tokenized funds is costly and is not competitive with LPs on IRR. It might be impossible to provide enough information to make a private VC fund buyable by traders. US rules recognize this and block most trading.
The LP structure, which is now 50 years old, also has some big advantages that will make it hard to push aside. It is popular because it is investor friendly. Investors know they are going to get all of their money back. The funds are “self liquidating”, with all of the money from exits being distributed back to partners. Investors are buying a package with well defined tax reporting, tax impact, and no double taxation. They can negotiate a fair division of fees and expenses between the investors and the GP. Their managers are incentivized to return cash before taking their fees.
Tokenized funds bring more unknowns.
Using the money — disclosure
Investors want to know how their investments are doing, when they are likely to get money, and what the underlying assets are worth. This is especially important if they are buying and selling those investments. VCs cannot provide very much of this information to the public. They often know a lot about exit plans and potential exits, but they need to keep this information confidential during the long process of arranging a deal. This secrecy is a big problem for funds that are packaged as securities. It prevents them from being actively traded.
Getting the money — uninvested funds
An LP fund increases measured profitability (IRR) by making capital calls only when they are ready to make an investment in a portfolio company. Then they immediately use the money to make 20% average returns. That way, they don’t hold uninvested funds in a bank account, which would have a 2% return. A fund that sells securities before they make investments will be holding uninvested funds, and it will have a lower return. The first generation of tokenized funds has tended to receive and hold their funds as crypto — bitcoin and Ether. In the first year, most of the performance of these funds is determined by the price of bitcoin. So, the buyers of these funds are getting bitcoin, but with high management fees, and much less liquidity. That’s obviously not a competitive investment structure, and I don’t expect it to continue as a best practice.
Getting the money — sales costs
An LP-style VC fund will allocate about 2.5% to sales costs, and pay for this partly with management fees. This is an acceptable cost to investors, who are concerned about every percentage point of return. In theory, selling a tokenized and securitized fund could be even cheaper, because the fund can be sold from a Web portal with a few forms. Fundraising is often more expensive in practice. Sellers pay marketing costs to reach investors on roadshows and through media, or they are pay sales costs to brokers. This reduces investor ROI.
Returning the money — distributions and buybacks
After an exit, an LP-style fund simply sends every investor their share of the money in a “pro-rata distribution”. Most of the first generation of tokenized funds instead use a buyback. They use the money from an exit to buy back the tokens. Investors get money by selling some tokens, and they hold a smaller group of remaining tokens to match the smaller remaining portfolio. Buybacks are simple for the fund because they do not have to track down their investors and send them money. And, theoretically they are simple for the investor, who is already equipped to sell securities and record a capital gain or loss, and does not have to do other types of income reporting (unless they are in the US).
There are at least two ways to do buybacks. In a proportional buyback, the fund buys a certain percentage from each investor. This is almost the same as doing a pre-rata distribution. In a market-priced buyback, the fund just goes to an exchange and buys tokens. This is very different from a distribution. In theory, it creates economic value, because the investors get the value of the option to stay in, or exit. Some of them will stay, and some of them will sell, and both will feel like they have more value than they got from the proportional distribution. And, it creates liquidity, which is also useful for investors. However, it relies on the idea that investors can accurately estimate the value of the remaining portfolio. Valuation could be a lot of work. If they can’t do that, then their decision about buying or selling will be unfair.
Some of the new funds are “evergreen”, which means that they do not give the money back, or they only give half the money back. This allows the managers to continue collecting fees for many years, and this type of fund often trades at a discount to its asset value.
No trades please, we’re American
A traditional long-term limited partnership flourishes in the US, under US rules. However, the SEC and the IRS have teamed up to make it difficult to sell and trade private funds as securities. For this reason, many first generation tokenized funds do not accept investors from the US.
A fund with US investors has limits on the number and type of shareholders. A private fund becomes a “reporting” public fund, with a lot more regulatory obligations, once it reaches 100 accredited investors. That is not enough to create a resale market. You can have 2000 investors if you exclude accredited investors and only accept “qualified purchasers” who have at least $5M to invest.
If a fund organized in the US has more than 100 partners, and it is “readily tradable on a secondary market” then the IRS will classify it as a corporation and start asking for corporate tax, which is bad. The company then has to either become a publicly traded partnership that distributes 90% of its income, or go offshore.
Funds that go offshore are less accessible to US investors. US investors in non-US funds need to be aware of the rules for “Passive Foreign Investment Companies”, or PFICs. The US government does not want taxpayers to invest in a foreign fund, and then compound their gains tax free. So, they apply heavy tax penalties when you do realize those gains. They apply a high tax rate, and then they ask the taxpayer to pay interest on taxes that might have been due from previous years. Very bad. Taxpayers can avoid this fate by getting an annual profit and loss report from the foreign fund, and then filing a “QEF election” to pay taxes on gains every year.
The tokenized funds that I looked at are definitely PFICs. If they don’t commit to offering annual P&L reports, they will be unbuyable for US investors who submit accurate tax returns and are not tax exempt. I haven’t seen any funds that have warned their token buyers about this.
There are limits on how a US private fund can be exchanged. The basic idea is that exchange should be a slow process which allows months for due diligence, and this approach is codified in regulations.
It will take some work to include Americans in a tokenized fund market. I expect to see the following adaptations when selling to Americans.
Organize the fund outside of the US
This avoids the tax rules that fall on a “publicly traded fund” in the US. Caymans and Singapore are good options. A fund just needs to make sure that US investors can pay their taxes correctly. The fund will be truly global and it will have an easier time attracting non-US investors. The non-US fund also has more investment opportunities than many US funds, because it will not be banned from token sales and exchanges that exclude US investors.
Provide annual profit and loss reporting
American investors need annual P&L reports so that they can file a QEF election and avoid PFIC tax penalties.
Bundle investments into public funds
The shareholder and trading limits go away for “1940 Act reporting companies” — publicly traded BDCs (Business Development Companies), mutual funds and MLPs (Master Limited Partnerships”. They can have lots of trading, and they can have lots of shareholders, and those shareholders don’t need to be accredited or QP. These entities follow public company rules and file with the SEC as a “public fund” or a “reporting company under the 1940 Act”. VCs hate to do this because they feel that it creates a lot of extra compliance cost. Historically, these vehicles have not been competitive with the top private VCs, and there must be a reason for that. However, there are plenty of funds that go this route. When there are enough assets available in the tokenized security market, public funds will become an obvious channel for American investors to get into the market for tokenized private securities.
Raise at least $100M
A fund that manages more than $100M is a “Qualified Institutional Buyer” or QIB. QIBs (pronounced “quibs”) have special privileges in the US market for private securities, because of something called Rule 144a. They can buy and sell private securities amongst themselves without lockup. They can buy Reg D securities that are still under lockup. They can also buy and sell in the offshore “Reg S” market. That’s why our Aboveboard QIB whitelist is so awesome!
Offering sizes over $100M will be helpful in creating a liquid trading market.
Use Qualified Purchasers (QPs) and track them.
A US private fund can have up to 2000 participants if they are all “Qualified Purchasers” with more than $5M to invest. This rule is an adapter for international funds that want to include US investors. We are building a QP whitelist that will allow its members to participate in the international market for private funds.
Start with SPVs
Finally, we can eliminate the fund and just package a single investment into an LLC, which would be called a “special purpose vehicle” or SPV. This isn’t regulated as a fund, and it fits into the normal market for private securities. I will cover this option in the next section.
Ideas and options
Special Purpose Vehicles and the unbundled VC
Instead of selling 30% of a startup company to an established investment fund and making a capital call, we could package the investment into new company (usually an LLC) that just owns the 30%. We could call all of our friends and ask them to buy units of the LLC, so we can turn around and buy the target investment. We tell our friends we’ll take care of managing the investment to make money, and ask for a slice of the profits. This “Special Purpose Vehicle” or SPV is just a normal private partnership, not a fund. And, it can raise money in new markets, such as the less regulated token markets. This type of SPV is very common. It is often used to buy buildings and real estate developments. Groups like AngelList have attempted to organize ecosystems to do VC deals this way.
Fundamentally, a VC is a service provider that investors hire to scout for deals, negotiate investments, and manage them to exits. Investors can buy this service on a deal by deal basis using SPVs. They could pay a fee to get into a deal that a VC initiates. Then, they would pay the VC a share of profits on exit. This would be an unbundled VC model.
The problem with this strategy is that it chops the assets into small deals that are difficult to trade. Maybe we need to bundle them up into something bigger.
Bigger is better if you want to find other people to trade with. You will find a more buyers and sellers for bigger issues. If VC funds want to give their investors a resale market, they will bundle a lot of fundraising into a few securities.
Bundling fits together with SPVs. We create an SPV to unbundle the investment origination and servicing role the way we have unbundled the origination of mortgages. Instead of doing it inside a whole fund, originators can create these deals, and take responsibility for them, and fund them on a deal by deal basis. The money can come from funds that put money into a portfolio of these SPVs. The funds will be bigger, and have manageable risk profiles, and will actually trade.
Funds could be redeemable. Investors could ask for their share of the cash and securities in the fund, and walk away. This system would use the tools of tokenized or “programmable” securities. The portfolio securities could be programmed so they are transferable even during the lockup period for sale, and in this case governance rights might be proxied to the fund. VC and PE funds would overlap with hedge funds, which often allow redemptions each quarter.
It costs $1M to organize a $100M fund and start the sales process. The crypto world would fund that with a “presale”, and give the people that put in the first $1M a double share (a “100% bonus” or a “50% discount”). This costs everyone else an extra 1%, but it would be worth it.
I saw a VC fund that was offering a 30% bonus to the buyers of the first $10M. This is obviously a bad idea, as nobody would rationally buy the 10M+1 unit and pay that cost. But, paying 1% to get more professional packaging sounds like a good idea.
Funds that sell securities, and then store the proceeds in a bank account, will have lower IRR than limited partnerships which make a capital call only when they have a hot investment opportunity. Can we use the capital call system with security owners? We might ask security buyers to “stake” 20% of the portfolio value on initial purchase. Later, the security buyers would face a choice of delivering the money for a capital call, or giving up their 20% stake.
Tokenized funds are promising that their investors will be able to trade. However, it is difficult to price and trade a private fund because of the confidentiality and lack of disclosure that VCs engage in. What are some tactics for solving this problem?
Audits: Funds with outside investors will hire an auditor to check the books, and check the methodology for calculating the Net Asset Value (NAV) that is reported every quarter. It’s probably also a good idea for funds to allow investors to elect a representative to look more closely at the audit.
Buybacks: Funds can buy back their tradable units if they think the price is too low. This can fix some the disclosure problem. Fund managers can signal when the price is too low with buybacks.
Royalties: Passive security buyers might prefer to get a percentage of revenue (royalties), rather than equity. The payout is more predictable, and a monthly royalty payment also serves as a revenue report.
Full transparency: It might be possible to run a fully transparent fund that blogs and tweets and reports on almost everything they are doing. This fund would have access to a limited number of deals which could tolerate full transparency. It might work as a container for open source crypto deals, which are continuously exposing information, and for smaller SPVs and semi-public securities that have similar transparency obligations.
Open and automatic transparency: Cryptocurrency projects can be actively traded because they are open. The evolution of the project is visible in real time in the open source code, the network of nodes on the Internet, and the activity on the public blockchain. A portfolio could be made up of projects with this type of automatic transparency.
Look to Real Estate
Real estate is a much bigger asset class than VC, and it will account for a bigger share of tokenized investment. Real estate is also a better candidate for trading, because it does not require much confidential information. We can look at the structures that real estate investors are using.
These typically include a lot of special purpose vehicles to buy specific developments, some funds that are basically private REITs, and some “under construction” development projects which are essentially venture capital for the developers.
In my observation, the first round of real estate token offers has the same problem that afflicts the rest of the token universe. The deals aren’t very good. They are overly risky compared with the returns, from unproven managers, with uncertain rights and limited liquidity. They aren’t competitive with publicly traded REITs on one side, and with traditional private investments on the other side. The solution is to offer better deals.
The specific problem with many real estate deals is risk. Instead of offering cash-earning properties, which are easy for outside investors to understand, they are mixing in a lot of “under construction” venture capital. In the real world, passive outside investors with no control rights and no seniority avoid construction projects. They are very risky. Even good developers go broke when there is a financial crisis, and then they cram down their investors and lenders, and start over. Often, investors will make the developers put their own money into the construction projects, and then give developers money for the next development by buying the completed developments. In the token world, developers seem to be mixing in a lot of construction projects, assuming that token buyers like risk, and somehow don’t have the same concerns as real investors.
Funds can help solve this problem. A fund would buy into SPVs where developers take more risk, or have finished development. A fund would also be able to create “tranches” that have lower risk because initial losses are absorbed by a high risk investor.
The exposed tokenized format will make it easier to monitor the behavior of these investments and prevent the problems that happened in 2008. In 2008, tranches of mortgage backed securities started taking losses that were apparently unexpected and previously invisible. This crippled the world financial system. There is a better way.
Reselling existing VC portfolios
After VC’s invest in a company, they face a long road to getting their money back. Investors want their money back, and managers are eager to move on to investing new funds. VCs often ask if our upgraded private security market will provide place for them to get money back by selling portfolio companies, or an entire portfolio. This “secondary” market for selling VC portfolios exists now, but it is quite expensive. Buyers expend a lot of energy assessing the value of the portfolio, and they demand a substantial discount. With standardized forms of disclosure, it might possible to package existing interests into a tokenized SPV and sell them. At the right market clearing price, tens of billions of dollars of these outstanding interests might change hands.
This excellent presentation from Blackstone gives us some data about the size and pricing of the market. They show about $2T of “remaining value” in PE funds, of which about 1/6 or about $300B is in VC funds.
They note that at the peak of the market in 2016, “the average price paid for venture capital assets, with less predictable cash flow streams and less perceived upside potential, … was 78% of NAV, compared to buyout pricing, which averaged 95% of NAV in the same year.” Buyout portfolios own mature businesses with predictable cash flows, and they sold for up to 95% of NAV. Their 5% discount is about the same as the cost of selling an illiquid new issue, so we can imagine it as a liquidity discount. The discount on a VC portfolio with unpredictable payouts was 22%. These numbers imply that we only get about a 5% boost from using a more liquid packaging, and we might find that securitized VC funds sell at a 17% discount to NAV.
SpiceVC was one of the first tokenized funds to come to market. They seem to have raised between $10M and $15M. I consider their terms to be quite fair.
- They return all of the money from exits (minus fees) in the form of buybacks. Buybacks are offered proportionally to each shareholder, so this is really a proportional system, and it will work even if shareholders do not have a way to estimate the value of the remaining portfolio.
- They take a 15% carry, after they have returned 100% of the originally invested amount.
- They take a 2.5% management fee, which is very fair considering that this is a small fund, and they seem to have rolled some of the sales and organizational costs into this fee.
- They take a 15% token allocation for management and affiliates. This is where things get tricky. An investor that puts in $100 only gets $85 worth of tokens. The NAV has to go up by 15% before the investor breaks even. In a normal LP fund, a big chunk of the fund does not go to management, and this allocation is 0%, and an investor starts out with close to 100% of their investment. They could argue that if they eventually get a 4X payback, the reduction in carry from 20% to 15% pays for this management share. So, this is really a way for management to cover the downside of putting time into a small deal.
- The incentives of the managers seem well aligned with the investors. The managers make money when investors make money, off their carry and tokens. Incentives are important, and later will we describe the Andra deal, where the incentives are very different.
- They excluded US investors. This is a UK and Israeli operation that makes some investments in the US, but in its initial issuance did not deal with the rules governing US investors.
- They are publishing quarterly NAV disclosures. I found their NAV report to be good reading, and it shows a profit. But, this report illustrates the problem with VC funds doing disclosure. They provide no details for their claim that the NAV is $1.308. There is no way to guess what investments (or even the largest category, “unused funds”) have had mark-to-market events. They list four portfolio companies, but they do not reveal if one of those companies is 1% or 30% of the NAV. So, it would be difficult to bid for their security even if you had an opinion about the portfolio companies. They note that “We, at SPiCE VC are trying to hold the stick at both ends: on one hand to create a transparent report for our investors, while on the other hand, respecting the portfolio companies’ need for keeping their performance and financial information private.”
Securitize.io was the first investment from SpiceVC and helped sell their deal, so these projects are partners. Securitize has since used their portal software to sell a number of different funds and fund structures, so they have more experience than most. In conversations, I like their thinking about how to achieve liquidity.
Harbor has positioned themselves to be a technology designer and compliance manager for the new generation of private funds. At an event in May I talked with their CEO and heard him pitch the potential of liquid funds as a way to invest in new asset classes like art, and real estate neighborhoods. However, since then, they seem to have … disappeared. Given their focus on compliance, they might have fallen into the black hole of US regulation described above. They raised $42M, with $14M last October, and $28M from top California VCs announced in April. Sophisticated investors put a lot of money into a company with an unknown business model, no customers, and limited governance rights (they sold a convertible). These California VCs must see tokenized funds as a big opportunity.
OpenFinance is the first security token exchange to open for business in the US, and they are starting by listing funds. They say they want funds from SpiceVC or Blockchain Capital to be the first listings. I like their exchange architecture. OpenFinance is building a distributed exchange with open source software. One obstacle to tokenized funds is the lack of exchanges, so we need the emerging cohort of exchanges like OpenFinance to succeed. However, as a US exchange they are dealing with a lot of the regulatory restrictions around private securities. Plus, they are dealing with an additional layer of regulation that governs exchanges. In the view of the SEC, every exchange needs components like a “clearing broker” which are irrelevant if crypto technology is applied properly, but baked into the rules anyway. This means that a US exchange will be less efficient than an offshore exchange, and it will be forced into less active trading. OpenFinance may become fabulously successful as an exchange for tokenized securities that are doing full public filings in the US. This market that doesn’t exist yet, but in the future we will design programmable public securities with interesting features. For private securities, we may need to turn to non-US exchanges and build around the Reg S / 144a market. I am currently seeking interviews with exchanges that have announced they want to support tokenized securities, in the US and outside of the US.
CityBlock is a sort of fund of funds. They have stated an intention to raise money with a tokenized offering, and then distribute it to local VCs in several cities. I don’t know much about this deal, but I like the fact that is designed to create liquidity. Instead of investing in a bunch of small VC funds, each of which is impossible to trade, you invest in one aggregated fund that has enough mass to be tradable. They then distribute money to a bunch of local managers. This (hopefully) solves both the problem of liquidity (requires a large float) and size (smaller funds can have better returns). Personally, I believe that this approach will be stronger if it is combined with the “unbundled” VC approach described above, so they only need to send money out when there is a real deal. Maybe that is what they are doing.
Andra Capital released news in April claiming that they had raised $500M for their tokenized fund. I have heard varying opinions about whether they have actually collected the money. If the news is true, it is a huge milestone for the industry. It is also a problem, because the terms that Andra published do not create an investor friendly foundation.
The basic idea is that their “Silicon Valley Coin” gives investors access to late stage tech startups (“follow on investments”) that are available to Silicon Valley venture capitalists, but not ordinary investors. They claim a “target IRR” of 30%. They say they can achieve this return by getting into the top 20% of VC fund performance. They will return money by using 50% of every exit to buy back tokens. They can also sell new tokens. So, this an an “evergreen” fund that continues indefinitely, rather than a self-liquidating LP fund.
Here is what their term sheet and other marketing materials seem to say.
Costs: Their annual management fee is 3%, rather than the typical 2%. They are likely to have significant sales costs, because they are doing online marketing, a roadshow, and selling through Wall Street brokers. The terms state that the fund will pay sales costs beyond the first $500K, with no estimate or cap. Managing sales costs is important for any fund because an investor will start out with a negative return — a NAV of $95 on a $100 investment if sales costs are 5%. Andra will take 20% of every exit, after hitting some (poorly explained) trigger for returning money. Since their stated policy is to return 50% of exit money to investors in the form of a buyback, the carry fee is really 28% of the distribution (20%/(50%+20%)), significantly more than the typical 20%. One might expect a fund with that structure to trade at a 28% discount to NAV after the carry comes into effect.
Incentives: The published terms appear to incentivize managers to run the largest possible fund, which might not maximize investor profits. They get paid 3% of total NAV every year, which is quite different from a fee based on the amount that was originally invested. They appear to have the option of keeping the exit money and not distributing it with a buyback. And, they have the option to sell new tokens, so net distributions might actually be negative. They get 20% of every exit after a trigger, rather than 20% of profits on the exit. Under these rules, a manager might not return any money, and then extract the whole value of the fund 20% at a time by churning through exits. I’m not saying that anyone would do that, but it seems like it could fit the rules of the game.
Rights: If the incentives don’t work, and investors think they are being treated unfairly by a normal LP fund, they can make a legal complaint. However, under the terms of the Andra deal, investors have no obvious legal rights, because they don’t actually buy securities from the fund. Instead, they buy tokens from a shell company in the British Virgin Islands, and that company sends money to the fund in the Cayman Islands.
Claims: Andra publicizes a “target IRR” of 30%. They say they can achieve this if they get into the top 20% of VC fund returns. They are claiming they can get returns that are better than 4 out of 5 VCs, while their stated policy is to invest in the same deals that other VCs are investing in. A cynical observer might note that earlier stage VCs are unlikely to open their very best deals to follow-on investors. Under the US disclosure rules for public funds, a fund would have to list the reasons why this comparison to the top 20% of VC funds might not be exact. If they followed that rule, they would have to cite some structural differences that drag down returns. Costs are about 50% higher across the board (sales, annual, carry). They will be holding uninvested funds. They don’t actually give all of the exit money back, and historically this type of closed-end fund has traded at a discount.
I sent Andra a message with my concerns and questions, and I received this beautiful non-answer: “Thank you for the tremendous amount of time spent in reviewing our documents. Each of the points are thoroughly explained to each of our LPs during the diligence process. We’ve designed a friendly LP structure with our partners at DLA, Deloitte and advisors but our practice is to only educate and provide detailed information to serious investors. Most of your observations have been addressed and we are interpreting many of the points differently.”
My favorite part is the name drop of DLA Piper (their lawyers) and Deloitte (their auditors). Issuers often assume that they can make their deal look better to investors if they hire well-known law firms. However, those lawyers work for the issuer, not the investor. If the issuer wants to do something straightforward, those guys will make it straightforward. If the issuer wants to exploit investors, those guys will figure out a way to reduce the legal risk. Who do you think came up with the idea of selling tokens out of a shell company in the Virgin Islands? DLA Piper told me that they charge $300K to $400K to help set up a tokenized fund. The issuer has to lift that money out of someone’s wallet.
Some assembly required
The goal of an investment fund is to make money for the investors. The traditional LP structure has been optimized for investors. The first generation of tokenized funds has some disadvantages in delivering returns. We’re running a workshop process to design a second generation of tokenized funds that will be optimized for returns, and include some new advantages, such as resale. This article describes a bag of parts. Now we need to assemble them. Please clap, and comment here with your ideas, or chat with us on Telegram.