view

If Everyone Loses Money Together It's Fine

Also ping pools, Bitcoin 10,000, blockchain by other names, state taxes and non-GAAP accounting.

Steinhoff fallout.

Last week I emitted an appreciative whistle about the JPMorgan Chase & Co. banker who signed up for a chunk of the Steinhoff International Holdings margin loan, lost $143 million on the trade, and then got that trade mentioned in JPMorgan's earnings release. "You have to do something really special to get called out individually in an earnings release," I said. So double kudos to the Bank of America Corp. banker who also took down a Steinhoff loan, lost $292 million on it, and got called out twice in BofA's earnings. "The provision for credit losses increased $154 million to $162 million, reflecting Global Markets' portion of a single-name non-U.S. commercial charge-off," reported the Global Markets division, while the Global Banking division chimed in to add "Provision for credit losses increased $119 million to $132 million, driven by Global Banking's portion of a single-name non-U.S. commercial charge-off." (That may not refer entirely to the Steinhoff margin loan that toasted JPMorgan; Bank of America has other loans out to Steinhoff.) It's less embarrassing when you parcel the loss out among two divisions. Citigroup apparently did the same:

On Tuesday, Citigroup Inc. CFO John Gerspach refused to identify which client was behind a “single client event” that triggered a derivatives loss of $130 million in its equities-trading unit. He told journalists the same client issue was responsible for roughly 90 percent of its institutional clients group credit loss of $267 million. That client was also Steinhoff, according to a person briefed on the results.

Goldman Sachs Group Inc. is cagier, and makes it harder to back out a number, but it also seems to have felt the need to slip Steinhoff into its earnings release: Its Investing & Lending division's "Net revenues in debt securities and loans were $449 million, 2% lower than the fourth quarter of 2016, due to higher provisions for loan losses, primarily reflecting an impairment of a secured loan, largely offset by higher net interest income." It is also of course less embarrassing when you parcel the loss out among ... well, four banks so far, but there will be more.

Frankly at this point if you're not calling out a nine-digit Steinhoff margin-loan loss in your earnings release, that will be a little embarrassing. Why didn't you do this trade that everyone else did? Were you not important enough to get the call? That's not a great look for your margin-loan business; if you're going to be an important bank you need to get the calls on the big trades. Or were you too conservative to do a trade that Goldman and JPMorgan and Citi and Bank of America were comfortable with? That's even worse. Sure they were wrong, but that's not the point. The chief financial officer of the next customer looking to do a transaction with a bit of hair on it doesn't care about how smart you are; she cares about how reasonable you are, how likely you are to actually get the trade done. If you can't get there when everyone else can -- if your due diligence is too stringent or your committees are too ornery or your lawyers are too careful -- then you are not going to get a look at the next trade.

I feel stupider for having typed that paragraph but it is a real (though implicit) pressure on bankers. The best thing you can do, as a banker, is execute a lot of good trades. The next best thing you can do is ... probably ... execute a lot of trades? From first principles you might think that executing fewer trades, but only good ones, would be preferable, but that rarely seems to be the mindset. If you do all the big trades, your bosses and clients can see that you are important and busy and normal. If you turn down the big trades that everyone else makes, you look weird and nervous and possibly lazy. The pressure to get there on this trade, to book revenue today, has a tendency to outweigh the worry about booking losses later. If you say no now, you'll be an outlier, and have to answer a lot of uncomfortable questions. If you say yes and it all goes wrong later, at least you'll be in good company.

Anyway here are Goldman's earnings release, and Bank of America's release, presentation and supplement.

Ping pools.

Here is a story from Bloomberg's Annie Massa about "ping pools," which are basically a thing where if you want to buy a stock, you call up a big dealer and ask if they want to sell it to you. I mean, not quite. For one thing, you don't use the phone; it's all fast and electronic. (You "ping" them.) For another thing, by "a big dealer" I mean a proprietary high-frequency trading firm -- "Virtu Financial Inc. and Citadel Securities operate the largest ping pools," and Jane Street Group LLC and IMC Financial Markets also have them -- rather than a traditional investment-bank broker-dealer. Also you don't ping the dealer; your broker pings them on your behalf.

Still the market structure here is, schematically, if you want to buy a thing you call up a big dealer and see if they want to sell it to you from inventory at a competitive price. This is a very traditional market structure. It is roughly speaking how the art market works, or the market for cars; it even looks a little bit like how a supermarket works. (You don't call the supermarket, but if you walk in and don't like the price or selection you can go to another supermarket and trade against its inventory.) It is also pretty close to the traditional structure of over-the-counter markets like the bond market, where investors call big dealers individually for a price, and the dealers trade with the customers out of their own inventory.

On the other hand it is very different from the traditional structure of the equity market, where everyone is supposed to trade on an equal basis using orders posted on public exchanges. And the exchanges don't like it:

“All forms of dark trading should embrace transparency,” said Bryan Harkins, head of U.S. markets and global foreign exchange at Cboe Global Markets Inc., among the U.S. exchange operators that stand to lose business.

Otherwise, he said, over time it could undermine the best price in the open markets, a measure known as the National Best Bid and Offer.

“Choice and competition is good. However, the sanctity of the NBBO is paramount,” Harkins added. “Anything that erodes the NBBO should be seen as worrisome.”

There is a widespread assumption that bond trading will increasingly look like equity trading: There will be more electronic trading, more all-to-all trading platforms (like exchanges), and less intermediation by dealers who trade as principals with their own large inventory. So it is always interesting to look out for places where the equity markets are moving the other way. All-to-all exchanges with public prices have a lot of benefits, but some people prefer a more old-school approach even in equities. Customers might like the certainty of execution that they get from trading against a dealer's inventory instead of having to look for a natural counterparty; dealers like knowing who their counterparty is to minimize adverse selection. It would be weird if bond or derivatives markets move entirely to exchange-based trading, given that equities still resist it.

Happy Bitcoin 10,000 Day!

"Bitcoin tumbled below $10,000," this morning, "bringing its loss to almost 50 percent from a record set only a month ago," so that is a fun reverse milestone. I should get "Bitcoin Zero Before Dow 30,0000" hats made just in case. If you bought bitcoin a month ago, you've lost almost half of your money, but on the other hand, if you bought bitcoin a year ago you're still up about 1,000 percent. We talk sometimes about how cryptocurrency involves re-learning the lessons of financial history in a sped-up way; perhaps it also involves reliving financial history on fast-forward. Maybe last January was the Industrial Revolution, the last couple of months were the Gilded Age and the Roaring '20s, and yesterday was the crash of 1929. The rest of this week will be the Great Depression, and Crypto World War II will break out over the weekend. I am looking forward to reliving the 1980s in February, and maybe the dot-com boom in early spring. Anyway, here's "Did Bitcoin Just Burst? How It Compares to History's Big Bubbles."

Transformational data sharing transformational data sharing transformational data sharing.

Ahahahahahahahahaahahahahahahahahaha:

Two ETFs slated to launch Wednesday had to erase the word “blockchain” from their name after a last-minute intervention by the Securities and Exchange Commission, according to two ETF firms. The SEC insisted on the name changes amid a speculative frenzy for anything associated with bitcoin or blockchain, the technology behind the cryptocurrency. ...

As a result of the SEC’s prohibition, the Reality Shares Nasdaq Blockchain Economy ETF will be called the Reality Shares NexGen Economy ETF and the Amplify Blockchain Leaders ETF will instead be named the Amplify Transformational Data Sharing ETF. The former will trade under the ticker BLCN, while the latter will use the symbol BLOK.

At least the tickers can still be blocky. The exchange-traded funds will buy shares in companies that do blockchain-y stuff, but not exclusively blockchain-y stuff (think IBM and JPMorgan, not Long Blockchain Corp.), and the worry is that if you so much as whisper the word "blockchain" around retail investors they lose their minds. This is a particular problem for an ETF, whose price is supposed to track the prices of the underlying securities. If a small blockchain ETF debuts and immediately goes up 500 percent, while the underlying stocks -- huge stable stocks like IBM and JPMorgan -- don't move, then you have a bit of a problem.

Elsewhere:

Maersk, which owns the world’s biggest container operator, said Tuesday that it will enter a joint venture with International Business Machines Corp. to create a more efficient and secure platform for organizing global trade using blockchain technology.

They also announced this last March, but I guess these days you can never announce your blockchain initiatives too many times. What if someone forgets that you blockchain blockchain blockchain? That's why the blockchain name change is so effective. No one would forget that International Blockchain Machines is blockchaining blockchain blockchains to solve the global blockchain blockchain of blockchains.

New York may get rid of its income tax.

I am looking forward to living in a low-tax state I guess:

New York state would end income taxes on wage earners and make up the revenue with an employer payroll tax that’s federally deductible as part of a restructuring plan that Governor Andrew Cuomo is recommending to mitigate harmful effects of the new U.S. tax code.

I don't know if this will work, though it is probably sounder than the alternative approach of asking for charitable donations to the state and crediting the donations against the state income tax. Mostly I just like that:

  1. One of Congress's purposes in eliminating the state and local tax deduction was to encourage states to lower their taxes.
  2. Eliminating the state and local tax deduction will cause New York state to lower its income taxes to zero.
  3. But by cheating.

I love it when the system works, but in the opposite way to how it was intended to work.

People are worried about non-GAAP accounting.

Yesterday Citigroup Inc. announced the largest quarterly loss in its history, $18.3 billion, leaving it with a net loss of $6.2 billion for 2017. The stock was up a bit. I keep waiting for someone to write an article about how Citigroup is using "fantasy numbers" to bamboozle investors. After all, while Citi reported massive losses under U.S. generally accepted accounting principles, it also reported "Adjusted Net Income" -- that infamous non-GAAP dodge in which companies "adjust" out all the things they dislike to give investors a rosier view of their finances -- of positive $3.7 billion for the quarter (and $15.8 billion for the year). Citigroup's real, hard, generally accepted numbers were terrible, but it also reported flattering fake, soft, non-GAAP numbers, and investors were so distracted by the non-GAAP numbers that they ignored the real ones.

But I don't think anyone has written that article yet because it would be stupid. Citigroup's fake numbers -- the "adjusted" ones -- were really the real ones. It had some revenue -- $8.4 billion in consumer banking, $8 billion from institutional clients, plus some miscellany for a total of $17.3 billion. It had some expenses ($10 billion or so), and some credit losses ($2 billion or so), leaving it with $5 billion or so of pre-tax net income.

And then it paid ... $23 billion of taxes? Hmm that doesn't sound right. And of course it isn't. Citi's "Provision for Income Taxes" isn't the amount of cash taxes it incurred on its income in the fourth quarter of 2017. Instead it "primarily reflected the impact of Tax Reform." About $19 billion of the charge was "due to re-measurement of [deferred tax assets] arising from reduction in the U.S. corporate tax rate and shift to territorial tax regime"; another $3 billion or so was "related to the deemed repatriation of unremitted earnings of foreign subsidiaries." That is, the bulk of the charge reflects the reduced value of Citi's deferred tax assets: Citi has large past losses, which it can use to offset future income, but since the tax rate on that future income will be lower, the value of those offsets will be less. This is a very, very trivial problem to have, the opposite of a problem really. If you have pre-tax income next year of $100, and you can deduct $50 of old losses against it, and the tax rate is 35 percent, then you will pay $17.50 of taxes. If the tax rate is changed to 21 percent, then you will pay $10.50 of taxes. Careful examination reveals that $10.50 is less than $17.50, which means that the new tax rate leaves you with more money. And yet for GAAP purposes you have to book a large lossthis year, to reflect the declining value of your future tax deductions. 

Citi's "Adjusted Net Income" number attempts to answer the question: How much money did Citi make this quarter? Citi's GAAP net income number ... does not. It mostly answers the question "how much did tax reform reduce the present value of Citi's future tax deductions due to its past losses," with a little bit of the quarter's actual performance thrown in to confuse matters. One of those questions is of considerable interest to investors who want to understand Citi's business and its future earnings power. The other question is a curious bit of tax-reform trivia. 

Again, this is all so obvious that no one is actually going around criticizing Citi for reporting adjusted non-GAAP numbers. Obviously the non-GAAP numbers, here, are the "real" ones. But that is not always the case, and the next time you read someone criticizing a company for reporting (or investors for relying on) adjusted non-GAAP "fantasy" numbers, it is worth thinking about Citi. GAAP income is not real. GAAP has its own concerns, which may or may not correspond to the interests of investors who want to know how companies are doing. It has some big advantages -- it is comparable among companies, it is standardized, it is harder to cheat than "adjusted" income is -- but it is not a transparent view of objective reality. Sometimes it goes out of its way to obscure that reality.

Things happen.

Judge Signals He Will Approve $290 Million Settlement Between Allergan Shareholders and Pershing, Valeant. Troubles Push GE to Consider a Breakup. Celgene in Talks to Buy Juno Therapeutics. Melrose targets GKN with largest firm UK hostile bid since 2009. Beware the $500 Billion Bond Exodus. Filings reveal DRW’s climb into top rank of trading firms. How a 22-Year-Old Discovered the Worst Chip Flaws in History. Cyberattack on Crypto Investors Came From North Korea: Report. CFPB to Reconsider Obama-Era Payday-Lending Rule. "Observed Federal Reserve rulemaking participation by publicly-traded banks accounts alone for $7 billion in excess returns in the post-Dodd Frank era." If SCOTUS finds SEC ALJ appointments unconstitutional, what happens to tainted cases? Awl Ends.

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 mlevine51@bloomberg.net

    To contact the editor responsible for this story:
    Philip Gray at philipgray@bloomberg.net

    Before it's here, it's on the Bloomberg Terminal.
    LEARN MORE
    Comments