Investing Robots and Goldman Deposits

Also Acorns, Argentina, lobbying, cartels, duration and non-GAAP accounting.

Artificial intelligence.

Here is a story about how robots may eventually replace humans in investing, and I have to say: Of course? I like to compare the rise of artificial intelligence in investing to the rise of artificial intelligence in cars. Driving a car requires real-time decision-making and subtle coordinated physical responses to a constant flux of unpredictable visual cues, and if you mess it up you might die. Investing is so easy that you can do a slightly better-than-average job without making any decisions at all: If you just buy all the stocks in the index and sit on them, you will outperform the majority of professional active stock-pickers most years. You can think of an index fund as an incredibly stupid robot stock-picker, and that incredibly stupid robot has already replaced human stock-pickers for trillions of dollars of assets under management, and gotten better results. Presumably slightly smarter robots -- "smart beta" robots, asset-allocating robo-advisers, etc., that look at factors other than raw index weightings -- can do even better, and actually smart robots -- with the deep learning and the neural networks and the terrifying singularity and so forth -- will be as obviously superior to humans for investing as cars are to horses for transportation. And now robots can more or less drive cars! 

Not you, though, you are much better than any computer, you will always have a job as an investment analyst, don't worry. Anyway here's Sudhir Nanda at T. Rowe Price:

“Having a human is still important. Humans aren’t going to be completely replaced, but they will be mostly replaced,” he said.

Elsewhere: Did you know that you can pay as much as 2.32 percent in annual expenses for an S&P 500 index fund? I find that sort of impressive, but my Bloomberg View colleague Barry Ritholtz is less charmed

Get your Goldman Sachs toaster.

One basic sensible shake-out that you might hope for from the financial crisis would be:

  1. Banks might fund themselves more with deposits.
  2. Non-banks might fund themselves less with deposits.

On the one hand, you might expect banking to become more boring, relying more on traditional well-understood safe-ish stable-ish sources of funds like retail deposits and less on flighty unstable wholesale funding. On the other hand, deposits aren't that stable, and the pre-crisis notion that you could fund whatever business you want with demand deposits that might be withdrawn at any time now seems a bit aggressive. So the question is just one of sorting; every bank-ish-looking thing has to decide if it's a bank or not. General Electric: Not a bank. Goldman Sachs: Maybe a bank? In any case it "last week launched, a platform it inherited via the acquisition of a $16bn book of deposits from GE Capital." Goldman's chief strategy officer, Stephen Scherr, called it "a different avenue to use, with a different orientation and a different tenor," from wholesale funding and brokered deposits. On the customer side, a "banking expert at NerdWallet" called it "synonymous with the dream of wealth in America," which would actually not be a bad slogan for Goldman's retail bank? Definitely better than "synonymous with blood-sucking cephalopods," which is probably the main marketing challenge. Disclosure: I used to work at Goldman Sachs -- both before and after it became a bank holding company -- and I am sort of contemplating opening up an account with the bank? Not only because "the six-month CD pays an annual percentage yield of 0.7 per cent, more than five times the national average," but also because, sure, I think there might be some cachet in banking with Goldman.

Elsewhere: "GE Money Bank, a unit of GE Capital in the Czech Republic, said on Monday that its initial public offering could value the banking business at 39.1 billion Czech korunas, or about $1.6 billion." GE Capital will keep a minority stake, and the unit will be renamed Moneta Money Bank, which is apparently not Czech for "Money Money Bank," though wouldn't that be great? Even just "Money Bank" feels redundant. Still elsewhere: "GE Uses Hot Sauce to Burnish Image With Potential Young Recruits," though that's about materials engineering, not financial engineering.


Here is the story of a financial startup called Acorns that lets its customers move "spare change" into exchange-traded funds "by rounding up purchases to the nearest dollar and depositing the difference." It has about 850,000 accounts with an average size of $175, that is, one hundred seventy-five dollars, and while on the one hand you have to start somewhere, on the other hand $175 will not fund much of a retirement. It's possible that the U.S. retirement savings system is broken. But, hey, maybe Acorns will fix it. This is ... promising?

As a DJ, CEO Noah Kerner he learned to “keep a crowd’s attention every night,” which he said more financial firms needed to understand to keep short-attention-span customers focused on a long-term product like investing.

To be fair, he also has experience as a startup founder and chief strategy officer at WeWork, but I like to imagine that the job interview focused just on the DJing experience. Anyway it is all so millennial-y that it might as well be written in emojis:

Acorns offers five levels of strategies ranging from conservative to aggressive, made up of six ETFs. Jeffrey Cruttenden jokes that it mimics the Tinder dating app: “Swipe right for more risk, left for less.”

It's Tinder for your bank account, but with this sick beat. Elsewhere: "How the Fintech Revolution Is Reshaping Our Economy." And: "Key Points from the OCC’s Financial Innovation Paper."


Here's a behind-the-scenes look into the final negotiations among Argentina, its holdout bondholders, and U.S. mediator Daniel Pollack, that led to a deal for Argentina to pay off the holdouts and return to the international capital markets. I don't quite know how to feel about this:

Mr. Pollack’s rule of no pens and paper during crucial negotiations frustrated some, with two hedge fund managers complaining privately that the mediator was an obstacle to the settlement process.

Is that a thing? Do they teach that in the Advanced Negotiations class at business school now? What problem does it solve? Is it about, like, mindfulness and presence and eye contact? I feel like I'd be furtively typing notes on my phone.  


Here is a good and funny Matt Yglesias article called "This is the real way big business peddles influence in Washington." It focuses on Google, and I would summarize the two "real ways" in which Google influences government as:

  1. It makes good arguments that persuade people in government that it is correct, and
  2. It makes good products that make it popular with people generally, so it is politically useful for government to be nice to Google.

Also sometimes omelets. The important point is: These are good ways to influence the government. If the government does things that are politically popular and justified by good arguments, that's a better-than-expected outcome, honestly.

Here is a story about how big U.S. banks are hiring economists to make cost-benefit arguments for rolling back some areas of financial regulation (or, "at the very least," "halting the march towards ever-higher capital requirements"). You may agree or disagree or be agnostic about the relative costs and benefits of higher capital requirements, but I submit to you that this approach addresses, at best, Real Way No. 1, making good arguments. Real Way No. 2, being popular with the public, seems pretty hopeless for the banks these days. And getting popular support for lower bank capital requirements, I mean, good luck with that. Elsewhere: "Campaign’s Populist Tone Rankles America’s CEOs."

The Cartel.

I am a little confused about how the Justice Department is thinking about prosecuting individuals for their participation in foreign exchange manipulation. Like, what the Justice Department has is:

  1. Logs of chat rooms with dumb names like "Cartel" and "Mafia" where bank traders talked openly about manipulating foreign exchange fixes.
  2. Those traders' trading records.

To me, those things either combine to prove a crime, or they don't. The whole thing is documented. I would think that anything else that the traders have to say would be superfluous, or perhaps exculpatory: If the documents don't prove a crime, then what will testimony add? But apparently the Justice Department is offering "reverse proffers" to some of the traders to try to get them to talk:

During them, the DoJ “says here’s our best evidence against you, and here’s why you will lose at trial,” said Roger Burlingame, a former US federal prosecutor who is now at law firm Kobre & Kim in London. “The government hopes the person will decide to plead guilty and co-operate to get a reduced sentence rather than fight the case.”

And I guess it worked a little:

A currency trader from the Cartel chatroom -- the instant-messaging group the U.S. government named in wringing guilty pleas from five global banks -- has been helping prosecutors who are trying to build foreign-exchange manipulation cases against individuals, according to two people familiar with the matter.

The cooperator, they said, is Matt Gardiner, a former UBS Group AG trader known in the chatroom as Fossil.

He "could bolster the U.S. case by helping prosecutors understand the market and actions by traders, and later explaining it all to a jury," perhaps starting with the nicknames.

People are worried about duration risk.

The thing about very low interest rates is that they ... [checks notes] ... yes, they make investing in bonds unattractive. The interest payments are so low! Also there is duration risk:

Investors are setting themselves up for damaging losses if average yields rise even a little from their rock-bottom levels. Based on a metric called duration, a half-percentage point increase would result in a loss of about $1.6 trillion in the global bond market, according to calculations based on data compiled by Bank of America Corp.

You can sort of reconstruct the math on that. Economic "gloom, combined with more aggressive stimulus measures by the Bank of Japan and the European Central Bank, pushed average yields on $48 trillion of debt securities in the BofA Merrill Lynch Global Broad Market Index to a record-low 1.29 percent this month, compared with 1.38 percent currently." At the same time, effective duration on the index "surged to an all-time high of 6.84 years." When rates go up, bond prices ... [checks notes again] ... go down; 6.84 years of duration times $48 trillion of debt gives you about $33 billion of principal losses per basis point of long-term rate increases. Ignoring convexity, that checks out to about $1.6 trillion of losses if rates go up by 50 basis points. Or put another way: If rates go up by about 20 basis points or so, the market as a whole will lose more on principal than it will make on interest in a year (1.38 percent of $48 trillion, or about $662 billion).  

“It takes a fairly small move out in rates on the long-end to wipe out your annual return,” said Thomas Wacker, the head of credit of the Chief Investment Office at UBS Wealth Management, which oversees $2 trillion in assets. Longer-maturity debt is “not something we are particularly keen on,” he said.

On the other hand: "As Bond Yields Rise, Some Investors Fear Another False Dawn." And with rates so low, Larry Summers argues that Harvard should spend less of its endowment: "If it makes sense for Harvard University to pay out 5 percent of its endowment in 1999 when the real interest rate was 4 percent, it’s really quite unlikely that it makes sense to pay out 5 percent of its endowment in 2016 when the real interest rate is zero." This cuts against the more popular view that schools with giant endowments should spend more of them, or else be taxed. It's a little microcosm of a broader economic debate: Low interest rates raise asset prices and make the rich richer, but the definition of being rich is not invading principal. And since their yields go down, they feel poorer.

People are worried about non-GAAP accounting.

So Company X reports income, under U.S. generally accepted accounting principles, of negative $100 million. But it also reports that, after excluding a non-recurring charge for Thing Y, its pro forma non-GAAP income was positive $50 million. There are three possibilities:

  1. GAAP, which is a man-made set of social conventions for how to present financial results, does not accurately reflect Company X's economic reality and continuing future earnings power, while the pro forma results do. (Or, less strongly, GAAP is informative, but the pro forma numbers are also informative, about a different aspect of the company's situation.) Investors know that it is a useful discipline to hold all companies to the GAAP social convention, but appreciate Company X's efforts to provide them with more insight into its real economic position, and use the non-GAAP numbers to inform their own models. 
  2. The GAAP numbers are more or less right, the non-GAAP numbers are just fake and flattering, but investors, whose job it is to make intelligent decisions about the future earnings power of the company, ignore the non-GAAP numbers and focus on the GAAP numbers, which are after all reported prominently in the same press release.
  3. The non-GAAP numbers are fake and flattering and all the investors are chumps who believe whatever management puts in the big font at the top of the press release.

Pretty much every discussion of non-GAAP accounting focuses exclusively on Possibility No. 3, which, as both an efficient-markets fundamentalist and an accounting post-modernist, I find by far the least plausible of the three possibilities. Anyway, here's a column by Gretchen Morgenson with the pleasingly archetypal title "Fantasy Math Is Helping Companies Spin Losses Into Profits."

People are worried about unicorns.

I am generally pretty laissez-faire about unicorn governance; if a company's founder can negotiate with sophisticated private-market investors to get herself millions of dollars in funding while retaining total control of her company, well, that's between her and them. And in fact that sort of governance-by-obsessive-tyrannical-founder can be very effective. It is possible, though, that it works better in essentially aesthetic activities like building a social network than it does in the heavily regulated, collaborative and scientific world of blood testing? Anyway, here's a story about how Theranos, the Blood Unicorn (Elasmotherium haimatos), despite having various fancy boards of directors and advisers, is under the near-total control of its founder Elizabeth Holmes. Here is -- perhaps literally -- the money quote, about the board of directors:

“You’re practically a paper tiger and have fiduciary responsibility,” said Charles Elson, a corporate governance expert at the University of Delaware.

It is not the best combination to negotiate for yourself. Elsewhere, here are some claims that Holmes is secretly really good at public relations.

Elsewhere in troubled unicorns, here is a story about the rise and fall of Powa Technologies -- a U.K. payments unicorn founded and run by ace salesman Dan Wagner -- that has this astounding, astounding passage:

In his only interview on Powa since the company’s collapse, Wagner told the BBC he was “bemused” by Powa’s collapse. He said: “It’s the business equivalent of walking across the street and being hit by a car. It’s one of those things that is completely random.”

“It doesn’t necessarily reflect on what we were building. It doesn’t necessarily reflect on the capability or the experience or the management capacity to deliver value. It doesn’t necessarily reflect whether the valuation was right or wrong. Certainly, in this case I can tell you it was just one of those extraordinary things that should never have happened.”

That is not a good lesson for a chief executive officer to learn from a bankruptcy. Is life meaningless and random? Maybe! But if you are going to run a company you have to act as if it isn't.

People are worried about bond market liquidity.

I don't know, the headline of that duration article was "It's Dangerous Out There in the Bond Market," surely that counts for something, right? Oh also, here is an article about how yields on high-yield energy bonds are "close to falling below a threshold that’s deemed distressed," which in turn is "making some investors worry that traders are ignoring risks that continue to rise in the market." That is: High-yield yields are not high enough to be alarming, which is itself alarming. It's not exactly about liquidity, but as an unfalsifiable worry it is hard to beat. 

Things happen.

Proposed Bank Pay Rules in U.S. May Boost Salaries, Deter Hiring. The Rise and Deadly Fall of Islamic State’s Oil Tycoon. Oil Producers Lock In Once-Snubbed Prices. Decades-Old Oil Dependency May Stymie Saudi Prince’s ‘Vision.’ The Bank of Japan's "exchange-traded fund purchases have made it a top 10 shareholder in about 90 percent of the Nikkei 225 Stock Average." Atlas Merchant Capital, Carlyle Group Team up to Bid for Barclays Africa Group. Warren Buffett’s Former Heir-Apparent Resurfaces as Activist Investor. Warren Buffett is "a friend (but not a relative)" of Jimmy Buffett. Billionaire Farmers Reap Fortunes From Russia Food Sanctions. SEC Settles With Apple Stock Day Trader. Did the American Dream die last week? Jackson Never Wanted to Be on the $20 Bill Anyway. Juice crawls. Italian dog-napping. Dog drives Tesla

If you'd like to get Money Stuff in handy e-mail 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.