Long Take: Why Peer-to-Peer models fail against oligopoly, with Lending Club shutting down p2p platform, Seedrs/Crowdcube merging, and Morgan Stanley buying Eaton Vance for $7B
Hi Fintech futurists --
This week, we look at:
Lending Club, the peer-to-peer lending innovator, turning off peer-to-peer lending after having a bank in its pocket
Consolidation of the UK's largest crowdfunders, CrowdCube and Seedrs, and their limited economics
The scale of the Morgan Stanley and Eaton Vance deal, creating a $1.2 trillion asset manager
The struggle of peer-to-peer models more generally, and whether the blockchain movement can overcome the Prisoner's Dilemma
How many Libertarians do you think there are in the United States?
Everyone, right? Everybody wants personal freedom and a limited government. Just listen to Twitter bots and the talking heads on the propaganda channels. Everybody votes their principles and is internally consistent in their logic. Long live Ayn Rand!
The answer is ... about 3% of the voting population.
About 3% of the population actually cares enough about their personal philosophy to lodge a particular vote in that direction. We could have picked on the Green Party instead, or any other policy-oriented group, and gotten the same result. The reality is that everyone else votes Democrat or Republican, because those are the teams that matter.
Everyone complains about Amazon, but we all shop online. We mourn the loss of the neighborhood coffee shop, but we buy Starbucks for the loyalty points. Thus the hypocrisy of human nature.
And here's the meat. We want peer-to-peer economies, grounded in our neighborhoods and tribes. We think that Wells Fargo, and Bank of America, and the Federal Reserve, and the rest of "them", whoever "they" are, are centralized monoliths running on papyrus and holding back innovation.
Right. Where do *you* bank exactly?
Do we really want peer-to-peer economies, though? Or are we lost in the poetry of utopia?
Remember Napster, and Kazaa, and BitTorrent, with their brick through the window of the media industry? Initially, the naive reaction of the labels was to build digital rights management into music players, and song files, and any teenager onto whom they could tattoo the letter of the law. DRM didn't work, right?
Certainly one way to look at the explosion of file sharing is to focus on the absolute figures of people consuming media for free. The core question there is to ask whether those people would be paying consumers in the market in the first place, or whether radio and mix-tapes have been replaced by the digital substitutes of "piracy", YouTube copyright infringement, and other modern artifacts.
We don't know the answer. We suspect, however, that if you had the patience to suffer through a DJ's advertisements or had the time to rip tapes, you might be the kind of person who has the capacity to deal with managing seeding and leeching. The clearest formulation on this topic comes in the article The Fifth Era of Recorded Music, from Bill Rosenblatt.
The media industry has been able to deploy a business model that uses the Internet to deliver a better user experience when bundled with the law. It is a worse user experience to avoid it. You can see how DRM-free downloads have collapsed as a commercial model in the chart above.
Put another way, a digital music company is as much a monopoly as its predecessor the record-label. Likely an even better one, given digital returns to scale. It is so good, that the peer-to-peer alternative loses as a value proposition.
In the same vein, it's hard to find good data on YouTube as a facilitator of copyright breach. But we know that a lot of websites and videos contain media content that a record label would otherwise try to monetize. If that media is not on Spotify, it is very likely on YouTube, accessible for free. A proxy for this content are the take-down requests under the DMCA reproduced below and numbering in the hundreds of millions.
Is that piracy? Maybe. It is certainly "file sharing". Is it peer-to-peer? Absolutely not.
Just because content is user-generated, that does not mean it is peer-to-peer. Google is the platform that mediates access and takes rent through advertising. Google is the platform worth over a trillion dollars today. And this realization takes us to Lending Club.
Lending Club represents an era of Fintech credit. The core premise at its founding was to recreate the dynamics of the sharing and social media revolutions. Instead of mediating everything through the centralizing machine of a bank -- and by the way banking licenses are sort of hard to find in 2006 -- why not create a connective platform like Kazaa? A bunch of people who need to borrow can show up with various credit risks. And a bunch of people who would like better investment returns can show up to assess those risks. And you, as the platform, take a cut.
Sound familiar? This section is a warning shot to Compound, Aave, and the rest of the DeFi protocols that think that redefining technology redefines market structure, human nature, and micro-economic behavior.
The first problem is getting good risks. If you are venue for emerging credit, the risks that come to your platform are subject to adverse selection and the lemons problem. So you need sufficient aggregation, correlated with heavy customer acquisition and branding costs, in order to create the asset class of reasonable credit exposure. This is also why digital asset fundraising platforms are having a hard time. Most good investment opportunities still want to raise money form Goldman Sachs, Google Ventures, and Andreessen Horowitz. Not Globacap -- despite such a site being a strong technical and market innovation.
The second problem is getting enough investors. Remember we started talking about Libertarians that actually vote their politics? The same dynamics are there for financial behavior. Nobody actually wants to do the homework of selecting Lending Club notes, unless they are finance geeks (like us) or have some adverse incentives (packaging the product for someone else downstream). And the investors you get, especially if they are retail, are lumpy and finnicky. Your liabilities do not match the time horizon of thousands of people, flickering about with their needs.
By the way, this is a problem Dimensional Fund Advisors solved forty years ago. Instead of selling their mutual funds to retail -- and dealing with constant redemptions and purchases -- they targeted only institutional distributors (RIAs). This strategy meant that the financial product had less turnover and generated better returns. It all worked, until the form-factor of ETFs made their product advantage irrelevant by shifting the cost of redemptions and purchases away through market structure.
So you soldier on and bring in hard-nosed hedge fund capital. A private equity firm here and there, to package up all those Lending Club notes and smooth out the risks. Maybe sell them downstream into fixed income funds. Of course the cost of this funding from alternative financiers is really hight, because their job is to take the entire economic return and you have no pricing power. So you decide to aggregate your own capital through deposits and buy Radius bank.
And then you give up on peer-to-peer lending entirely. You're a bank now anyway. Why would you need this onerous many-to-many platform, when you can just offer some "high-yield" savings accounts.
It sucks. Peer-to-peer lending is dead. It was never going to work without a centralizing function to standardize deposits and slice up the risks. And the amount of people who "want" peer-to-peer is like the number of Libertarians. You and I still bank at the financial incumbent for 80% of our needs, and send 5% into a Fintech digital lender for experimentation.
What's the exception? What's the Google of this world? Let's look at our friends in China.
This geography too had a p2p lending explosion, which in large part involved fraud and bankruptcy. From the peak of 3,500 digital lender platforms, around 600 remain standing. Among them are the giants of Ant Financial and Tencent's WeBank. The high tech platforms outlived all of the individual fintech competitors, and used their size and credibility with regulators to remain in business. Everyone else is being effectively shamed and shut down.
Returns to scale have come from being a technology monopoly. Financial features are the monetization cherry on top.
In an eerily similar fashion, the same challenge is hitting the equity crowdfunding industry. We have been bearish on these platforms because of the Libertarian (i.e., small market, low commitment) problem. The profile of a financial consumer that likes to make some-but-not-all financial decisions is a myth. The failures of Covestor, Motif, Kaching, and other active-ish digital wealth platforms in the US highlights the problem. The UK, on the other hand, still holds on to a functioning narrative about this sector.
Some of the early neobank players, like Monzo and Tandem, engaged with the crowdfunding market to raised single-digit million amounts from thousands of excited supporters. Those supporters were also early users of the neobank products. The positive relationship between investors and users spun out into the story that crowdfunding is a successful economic arrangement, and that the crowdfunding platforms themselves will be the next generation of investment banking. To do this, the platforms had to do the heavy lifting to impact regulation that created operating models allowing regular people to access the venture asset class.
And yet last week, Crowdcube and Seedrs (the two arch-rivals of equity crowdfunding in the UK) announced a 60-40 merger and a likely need for future growth equity. Here are their economics.
Seedrs is making £4.2 million and burning £4.6 million. Crowdcube is making £7.6 million and burning £2.5 million. Together, that's roughly £12 million of revenue and £7 million of loss. If revenue doubles, we expect the combined entity to be break-even. Together, the firms have put 1,500 startup projects through the fund-raising process and raised £2 billion since 2011. Assuming 200 startups per year, we are looking at £50,000 of revenue per startup. As a reminder, Goldman prints about $8 billion of revenue per year in its investment banking division.
It's depressing. Hard to paint the best case that is exciting for the "future of finance".
There are three take-aways for us. First, you have to make this market 1,000 times larger. If we were talking about a merger of £4 billion and £7 billion, it would matter a lot more. One way to do that is by bypassing the geographic and regulatory boundaries under which Seedrs and Crowdcube have had to operate. This is in large part why crypto markets print large numbers -- they are global, including the United States, Brazil, China, Russia, and the African continent. There is always demand somewhere.
Second, the adverse selection problems remain in the asset class. Why are these unique or exciting investment opportunities? Who really cares about putting money into a local small business and facing 100% loss, when you can buy Amazon stock and watch it go to $2 trillion? Who really cares about buying coffee from the local shop, when they have the Starbucks app and rewards cards? If you had more investors on Seedrs, would the Silicon Valley tech players (like Slack) decide to IPO there instead of the New York Stock Exchange? You can see this same theme playing out in the acquisition of SharesPost by Forge earlier this year.
And finally, there is hope. The incentive alignment between people who crowdfunded the neobanks and then became users of those applications is profound. This is exactly the dynamic that crypto protocols have been ideating around. See this write up: Liquidity Mining: A User-Centric Token Distribution Strategy, or the ConsenSys approach to the same problem here.
Crowdfunding works not when there is "access" but when there is something to achieve by participation. In today's world, that something is largely financial return. To be honest, it is sometimes confounding how Initial Coin Offerings -- the next generation version of crowdfunding -- were able to raise $20 billion over two years. Or how Decentralized Finance, the next generation version of blockchain-based capital markets, has been able to manage a $15 billion capital base.
Perhaps the capital itself is far more risk seeking, and is in the appropriate part of the portfolio (i.e., alternatives). Perhaps the community aspects are far stronger than in the crowdfunding model, and thus viral coefficients are higher, leading to faster social distribution. Perhaps the interoperability of issuance and trading allows for quicker monetization, and a sense that these markets are worth the trouble.
Or perhaps, unlike the media industry, the financial industry has not yet been able to deploy a business model that uses the Internet to deliver a better user experience when bundled with the law. We are all still working to figure it out.
Takeaways to Takeaway
We are in a world where Morgan Stanley has acquired Smith Barney, eTrade, and is now adding Eaton Vance for $7 billion. The esteemed institutional businesses are in the retail hen house.
That's $1.2 trillion in assets under management in manufacturing and $3.3 trillion of assets in distribution.
In the political sense, choosing between Morgan Stanley, JP Morgan, Bank of America, and Goldman Sachs is like choosing between Democrats and Republicans. Regardless of your niche political beliefs, you should pick a party that matters -- not the Libertarians. Don't take this as a comment on the current election, in which we can say the sane choice is far narrower (self-destruction vs. attempted redemption), but as a comment on power structure and how consumers of financial services behave.
Peer-to-peer models have not become a stable market equilibrium. While p2p activity continues in media, digital monopolies wielding the law have re-emerged and are more powerful than ever. In p2p lending, the original innovators have exited the business in favor of a more straightforward, scalable solution called banking. In p2p crowdfunding, the market is consolidating and showing limited growth economics.
Is this a feature or a bug?
What we can do in the blockchain experiment is to position mutually-owned protocols as market venues, and create feedback loops for both companies and users that incentivize them to choose open-source standards over closed solutions. But it won't be an easy win against human nature and our collective resistance to change. Linux and Wikipedia have shown us one way. Another way is that parts of the enterprise economy find meaningful value in decentralized networks and commit not to cheat in the Prisoner's Dilemma. Or perhaps, it will be a national priority for China to integrate all economic activity into its blockchain service network, and that will be the Sputnik moment for the rest of the world.
The answer is hard to know, but we have at least articulated the outlines of the question.