Money Stuff

Spotify's Non-IPO Really Is Novel

Also Fannie and Freddie, bitcoin suitability, digital currency, bribery and customer service.


Spotify AB has filed confidentially with the Securities and Exchange Commission to go public via a direct listing, in which it won't sell shares in an initial public offering but will instead just one day declare that it is public and let anyone who wants to trade its shares. We have talked a few times about this plan -- it's weird -- and each time I write about it someone emails to say "you forgot to mention that Google did something similar in 2004." That is ... not true?

Google's odd Dutch-auction IPO is often cited as a quasi-precedent for Spotify's direct listing, but it really isn't. First of all, Google did an initial public offering: It offered 14,142,135 shares (10 million times the square root of two, if you're following along with Google's math jokes) of brand-new stock from Google directly to the public, along with 5,462,917 more shares being sold by insiders, including Google's co-founders, employees and early investors. At the offering price of $85 per share, that was a total of about $1.67 billion, and represented 7.2 percent of Google's stock. Unlike Spotify's plan to just one day be public without raising money, and let insiders sell whenever they felt like it at whatever price the market will bear, Google issued new shares and coordinated with insiders to all sell a huge chunk of shares all at once.

They went further in coordinating: Google and its insiders also agreed not to sell for at least three to six months after the IPO. This was a traditional single-point-in-time IPO: Everyone sold all at once, and then cut off supply for a long time. Underwriters ask for these sorts of lockup agreements because they think that the overhang of additional insider selling is bad for the IPO, and Google agreed: Bringing together buyers and sellers into one big by-appointment event, with no new shares for months before or after, makes for a tidier going-public process. In a direct listing, by contrast, insiders can sell whenever they want: right when the company goes public, or when they wake up an hour later, or when the price settles down a bit a week later, or whatever.

And Google did not cut out Wall Street underwriters. Its deal was led by Morgan Stanley and Credit Suisse First Boston; eight other banks were on the cover of the prospectus, and 18 more were listed in the back. The underwriters were paid a 2.8 percent fee, which is pretty good for a huge offering. (Eight years later Facebook's IPO paid just 1.1 percent to the underwriters.) In a direct listing, you can cut out banks entirely if you want, though as a former capital markets banker I am tempted to say you should hire a bank or two anyway just for peace of mind.

So Google's IPO was in all ways absolutely normal, with one smallish exception: Instead of the underwriters building a book of demand at a set of prices and then vaguely describing the book to Google before getting the company to agree on an IPO price, Google ran an actual auction in which bidders were asked to submit their highest bids, and then the clearing price was calculated based on those bids. Even there Google made a concession to tradition:

We intend to use the auction clearing price to determine the initial public offering price and, therefore, to set an initial public offering price that is equal to the clearing price. However, we and our underwriters have discretion to set the initial public offering price below the auction clearing price. We may do this in an effort to achieve a broader distribution of our Class A common stock or to potentially reduce the downward price volatility in the trading price of our shares in the period shortly following our offering relative to what would be experienced if the initial public offering price were set at the auction clearing price.

Every initial public offering is, loosely speaking, an auction process: The banks go out to investors and ask how much they want at what prices, and the investors tell them, and then the bankers figure out what price will clear the market and give the shares to investors at that price. But only very loosely speaking: The price is set not at the highest possible price where the banks can sell all the shares, but at the (usually lower) price where they think the stock will trade well, and allocations are given not just to the highest bidders but based on who the banks and issuer think "should" get the shares. Google made the process more explicit and algorithmic, but it still reserved the right to adjust things in the old-fashioned way.

Spotify, on the other hand, oh man. "Without underwriters, Spotify shares won’t debut with a price based on investor feedback, with buyers lined up." Every initial public offering is really two auction processes. First there is the initial public offering, in which banks take orders and set prices and allocate shares. But then there is the opening of the IPO, in which people who didn't get shares in the IPO try to buy shares on the stock exchange, and people who did get shares in the IPO try to flip them for a quick profit. The opening of trading on the exchange is also done through a two-sided auction, coordinated by the stock exchange, just as every stock on the exchange opens for trading every day by means of an opening auction. Usually the daily opening auctions are uneventful, since everyone has a pretty good idea of what the stock is worth (because it was just trading yesterday afternoon). Opening auctions for IPOs, though, can be more exciting -- Facebook's, famously, was a mess -- because the IPO price is not always a reliable indicator of what the market price will be. And Spotify's opening auction -- with no previous trading price, no IPO price established by a pseudo-auction, a lot of interest from potential buyers of a hot new unicorn, and no coordinated supply from selling shareholders -- will be a whole new kind of fun. 

Fannie and Freddie.

"Fannie-Freddie Overhaul Might Mint Hedge Fund Riches, Or Losses," is the headline here, which seems about right. One thing that I have mentioned about the future of Fannie Mae and Freddie Mac is that it involves two unrelated questions, which are:

  1. What is the future of Fannie Mae and Freddie Mac?, and
  2. Should the government give billions of dollars to people who own Fannie and Freddie's common and/or preferred stock?

Both of those are hard questions! The first implicates difficult issues of market design and housing policy and systemically important financial institutions. The second is difficult because, when the government originally bailed out Fannie and Freddie in 2008, it did not fully nationalize them and left some stock in the hands of private holders -- and then later, when Fannie and Freddie were getting back on their feet, the government unilaterally changed the terms of the deal to zero the private shareholders. This was arguably unfair to them, though arguably it would have been perfectly fair to zero them initially. (Also, of course, a lot of the current Fannie and Freddie shareholders are hedge funds who bought it well after the bailouts for pennies on the dollar.)

But the answer to question 1 has basically nothing to do with the answer to question 2. If you want Fannie and Freddie preserved in their current form, or broken into smaller pieces, or privately capitalized, or backstopped by the government, or dissolved entirely, or whatever, that is one choice. And if you want to give the common and preferred shareholders billions of dollars (of cash or shares in the new entities), or no dollars, or something in between, that is another choice. There are some interactions between those choices -- if you are trying to raise private capital for New Fannie/Freddie, throwing a bone to the old capital providers might be helpful in building goodwill -- but they are analytically distinct. 


Whether shareholders make a killing or get wiped out might hinge on a yet-to-be written provision of a draft Senate bill that marks lawmakers’ latest attempt to overhaul Fannie and Freddie, which have been wards of the state for almost a decade.

The section currently reads “open pending further discussion,” said people familiar with the matter.

Makes sense. You can completely figure out how you want to overhaul Fannie and Freddie while leaving the shareholder stuff blank.

Bitcoin suitability.

Should big investment banks and brokerages help their clients buy bitcoin? On the one hand, bitcoin has gone up a lot, and if the brokers had helped their clients buy bitcoins a year ago those clients would be much richer today, and if your broker isn't making you richer then what is she doing? On the other hand, everything: Bitcoin is wildly volatile, it (mostly) trades outside of regulated U.S. securities markets, it has historically been linked with all sorts of criminal activity, bitcoin holdings get hacked all the time, and the fact that it has gone up a lot so far does not mean that it will continue going up. Also the philosophical purpose of bitcoin is to some extent to disintermediate the existing financial system, to allow people to control their own financial lives rather than having to rely on middlemen like Merrill Lynch, to free people from the tyranny of trusting their banks. That philosophical distrust is mutual -- Jamie Dimon memorably called bitcoin a "fraud" -- but it did start on bitcoin's side. Bitcoiners thought banks were a fraud long before bankers had ever heard of bitcoin. 

So this makes some sense:

Merrill Lynch has blocked clients and financial advisers who trade on their behalf from buying bitcoin, citing concerns over the cryptocurrency’s investment suitability.

The ban applies to all accounts and precludes the firm’s roughly 17,000 advisers not only from pitching bitcoin-related investments but also from executing client requests to trade the Grayscale Investment Trust bitcoin fund, according to a person familiar with the matter. The ban extends an existing policy barring access to newly launched bitcoin futures.

Here is the Financial Industry Regulatory Authority's rule on investment suitability, which requires brokers to analyze both "reasonable-basis suitability" -- that the investment "is suitable for at least some investors" -- and "customer-specific suitability." Merrill's blanket ban suggests that it has concluded that bitcoin is not suitable for any investors, which on its face is sort of a weird conclusion. Lots of very sophisticated investors have bought bitcoins and made tons of money doing it. (Lots of very unsophisticated investors have also bought bitcoins and made tons of money doing it.) 

But perhaps that is the wrong reading of Merrill's ban. Maybe instead Merrill has concluded that bitcoin is not suitable for any investors who would buy bitcoins through Merrill Lynch. If you want to set up a bitcoin wallet and deal with the exchanges, if you believe in the decentralized peer-to-peer nature of bitcoin, fine, whatever, go nuts. But if you just want your trusted securities broker to go buy you some of this newfangled bitcoin you've been hearing so much about, then perhaps you are not philosophically cut out for owning bitcoin.

Elsewhere in crypto-news, here is some more on Venezuela's nonsensical plan to create its own cryptocurrency linked to the government's oil holdings:

“The country is in a social crisis,” said José Ángel Álvarez, the head of a national association, Asonacrip, that has been working with the government on the Petro. “How do we manage to build trust? Open technology, clear rules that meet the attributes of the cryptocoin: decentralization, for example.”

Yes: Build trust through decentralization! Specifically, you give investors claims to oil owned by Venezuela's (central, untrustworthy) government, oil that might already be subject to lots of other claims from people who have been betrayed by that government -- unpaid lenders to the Venezuelan state oil company, for instance, or companies whose assets have been expropriated. If Venezuela had said "we are going to issue new bonds backed by our oil reserves," that would have been an obvious nonstarter, because of a combination of foreign sanctions, Venezuela's continuing failure to pay its old bonds, the existing encumbrances on its oil assets, and a general lack of trust in the government. But instead: the blockchain! The blockchain solves all problems! Can't fund the Venezuelan government due to sanctions? You can, on the blockchain. Don't trust Venezuela's government to pay you back? You should, on the blockchain. Don't think you can actually get your hands on Venezuela's oil? Smart contracts via crypto make this super doable. There is no limit to the number of dumb ideas that magically become brilliant if you just add "on the blockchain" at the end. 

And: "Is Your Startup Stalled? Pivot to Blockchain." Is your idea bad? Bad ideas on the blockchain are actually good!

Elsewhere in digital currency.

Here's a story about a country of 1.4 billion people that has quickly transitioned to digital currencies. It's just that the digital currencies are administered by Alipay and WeChat rather than by decentralized blockchains:

Though the U.S. saw $112 billion of mobile payments in 2016, by a Forrester Research estimate, such payments in China totaled $9 trillion, according to iResearch Consulting Group, a Chinese firm.

As with bitcoin, China's mobile-payments revolution started as a way to solve the problem of trust:

The path to mobile payment was blazed by Alibaba, which hosts online shopping bazaars where merchants sell goods to consumers. More than a dozen years ago, Alibaba, taking a page from the U.S. company now called PayPal Holdings Inc., started a system called Alipay as an escrow service. It would hold payments until shoppers received their goods.

The service caught on among buyers who didn’t always trust vendors to deliver as promised. 

You can solve that problem with a decentralized blockchain-based cryptocurrency and smart contracts, but that is sort of complicated and tedious. Or you can solve the problem by trusting one or two central intermediaries -- Alipay, WeChat -- to administer your trust in everyone else. Empirically, if you want to build an actual payments system, the old-fashioned solution of trusting a central intermediary seems to be a lot more efficient.

Everything is securities fraud.

On the one hand, yes, if a company's top executives take billions of dollars' worth of bribes -- "which allegedly included gifts of Rolex watches, $3,000 bottles of wine, yachts, helicopters and prostitutes" -- in a massive scandal that touches multiple successive presidents of its home country, that is definitely bad. And, yes, if the executives take those bribes in exchange for lucrative construction contracts that overcharge the company, then shareholders are clearly harmed by the scandal, and it would be nice if that harm could somehow be repaired.

Still it is a little weird that Petróleo Brasileiro SA, which is caught up "in a long-running investigation known as Operation Carwash," was sued for securities fraud by its U.S. shareholders (well, depositary-receipt-holders), and yesterday agreed to settle the case for $2.95 billion. As Petrobras itself points out:

In the agreement, Petrobras expressly denies liability. This reflects its status as a victim of the acts uncovered by Operation Car Wash, as recognized by Brazilian authorities including the Brazilian Supreme Court. As a victim of the scheme, Petrobras has already recovered R$1.475 billion in restitution in Brazil and will continue to pursue all available legal remedies from culpable companies and individuals.

It is clearly securities fraud -- in addition to whatever else it is -- to pay bribes and not tell your shareholders about it. It is also pretty clearly securities fraud to accept bribes and not tell your shareholders about it. But it is odd to think of this as securities fraud committed by the company: The bribes went to the executives, the company was harmed, and now the company has to pay out more money to make up for the harm. Petrobras's current shareholders -- who are ultimately the ones paying the $2.95 billion -- have been harmed twice. They should sue someone.

Elsewhere in alleged bribery scandals: "Michael Cohen, Formerly of Och-Ziff, Charged With Fraud by U.S."

Sorry you don't like our stock.

Here is a charming dumb story of investor relations: A guy on Twitter tweeted "Shorted @ocado and @boohoo again today ... both overvalued," meaning that he had sold grocery-delivery company Ocado Group Plc's stock short, and Ocado's customer-service Twitter replied "Oh no! We're sorry to hear this Mark. Please kindly DM us the email address and postcode on your account, in order for us to look into this further." It turns out Ocado's social-media team was just confused -- they thought he meant that he had been shorted, that is, that some items were missing from his delivery order -- but really that is a perfectly coherent response to his actual meaning. If someone tweets that he is short a company, why shouldn't the company (1) say "We're sorry to hear this" and then (2) ask for his home address? It's a nice mix of responding sympathetically to his concerns and being ever-so-slightly menacing.

Things happen.

Wall Street Aims to Thwart a Hacking Nightmare for Your 401(k). Intel, Microsoft Deal With Widespread Computer-Chip Weakness. Why Asia’s Booming Bond Market Doesn’t Need the U.S. Commodities Are on Their Longest Winning Streak in History. Tesla Breaks Another Production Promise With Crucial Model 3. In Uber-Didi War, Brazil Is Latest Battlefield. The Challenge for Vanguard’s New CEO: Keep a Behemoth Growing. Ray Dalio is worried. Turkish Banker Guilty in U.S. of Iran-Sanctions Conspiracy. Cash-Strapped Venezuela Offers to Pay for Medicines With Diamonds. Harvard Professor Says Foreign Militaries Could Help With Venezuela’s Regime Change. How's Gary Cohn doing? Slow smokers. Food-safety expert warns latest bizarre Silicon Valley $60 'raw water' trend could quickly turn deadly. What should Tyler Cowen ask me?

If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Thanks! 

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

    To contact the author of this story:
    Matt Levine at

    To contact the editor responsible for this story:
    James Greiff at

    Before it's here, it's on the Bloomberg Terminal.