Apple Fiction and Robot Hiring

Matt Levine is a Bloomberg View columnist. He was an editor of Dealbreaker, an investment banker at Goldman Sachs, a mergers and acquisitions lawyer at Wachtell, Lipton, Rosen & Katz and a clerk for the U.S. Court of Appeals for the Third Circuit.
Read More.
( Updated
)
a | A

Apple did a thing.

I mean, you know this. You can read about the new iPhone more or less anywhere, but for myself, I've moved past real reviews of real Apple products. For the iPhone 7, I have tried to restrict myself to reading fictional coverage. Fortunately there is a lot. You can get lists of fake iPhone features at the New Yorker ("Is Banksy"), the Onion ("Put together by a child") and Jezebel ("Call me when it transmits straight to my brain via Braintooth"). Fusion has a fake profile of the courageous product managers who took out the headphone jack. Here's a fake plug to get rid of the headphone jack in your iPhone 6. On Twitter -- the leading social network for surfacing fake news -- Sam Kriss reports that the new iPhone has a rotary blade, Jason Gilbert that it is "PRE-LOADED with 'Sully,'" and Kaleb Horton that its features "add up to a slow drip of morphine that allows you to ignore and dismiss the existential horror that confronts us all when we get out of bed." Also Bloomberg's Stephanie Davidson described the fake ant that comes with the new wireless ear buds. I am sure I am missing many others. Some of the real news sounds fake, too. Apparently the new iPhone will come in two different shades of black, one regular and one extra-scratchy.

But it is still not enough. My dream is to live in a world where I can read saturation coverage of imaginary Apple products, just like I can read saturation coverage of real Apple products today. I want whole publications that cover fake Apple news with the same obsessive fervor that real Apple news gets. I want our leading novelists to devote themselves to fake Apple reviews, but at the same time I want to never be quite sure whether I am reading a straightforward product review or the wildest inventions of a professional comedy writer.

Apple product launches are essentially science-fictional: Whatever happens to the headphone jack, each time Apple announces a new iPhone, it is giving us a new way to carry all of the information ever produced by humanity in our pockets, plus a better camera. It seems to me that the right way to honor that is by treating Apple product announcements as an opportunity to write our own science fiction, to try to come up with something even more outlandish than what Apple has done. I want the fictional coverage to crowd out the real coverage, and be confused with it, and replace it, and then to become real. I want Apple's reality to race with fiction, and then to pull ahead. Why not ants in the ear buds? Why shouldn't the feature list of the iPhone 9 just be "does whatever you want it to do"? What is to stop the characteristics of the iPhone from merging perfectly with each of our varied desires? The world will be Tlön. I will pay no attention to all this and go on refreshing my Twitter mentions. 

Personnel.

In this age of robots there are, loosely speaking, two ways to think about investing. One is to think about it: You have some ideas of what will work, some macroeconomic theories and fundamental analytical views and so forth, and you go invest based on those ideas, or program a computer to implement them. The other is to let computers think about it: You chuck a mass of data into a neural network, and let the network figure out what patterns are meaningful and predictive, and what to do about them. You don't give the network any preconceived ideas; it just looks at the data, and decides for itself.

The latter approach has some big advantages: The robot might well be smarter and more rational and less biased and emotional than you are. But it has a big weird disadvantage: You may not understand what the robot is up to. Like, all the signs that you can see might be telling you to buy stocks, and the robot is selling stocks, and you can't tell why it's selling stocks. And when you ask it why, it can't say something semi-comprehensible like "because stocks are rich on a trailing P/E basis" or "because the Hindenburg Omen was just triggered." Its answer will be of the form "look at all this data, man." The robot can find signals in the data, but the signals don't come packaged as human-readable rules of thumb with cutesy names and intuitive explanations and plausible narratives. The signals are just signals. They happen to work, but you don't know why. The robot somehow makes your world more complicated, alienating you from the underlying information.

Anyway I thought about all this when I read this Financial Times story about how Deutsche Bank and other companies are using online-dating-like algorithms to hire people. You fill out a computer survey, and your responses are compared with those of people who get good performance reviews at Deutsche Bank, and if you look like them then you're hired. The weird part is that there's no direct intentionality in what the companies are looking for:

Koru, the company that carries out the profiling, says there are no right or wrong answers, and no politically-correct judgments — only comparisons of candidate responses with those of top performers at the organisation doing the hiring. If those people happen to be difficult loners who never back down, so be it.

And:

Crucially, she points out that the most desirable behavioural profile, or “fingerprint”, is unique to every firm, and even varies within the same industry. Some fingerprints show clear ‘spikes’ in measurements of behavioural traits such as “grit”, “polish”, “teamwork” and “curiosity”. Others show a balance of different traits.

Nobody sits around saying "what sorts of people do we want to hire?" They just look at who they've already hired, and who has done well, and then hire more people like them. The data tells them the answer, and if it turns out that the people who do well in Deutsche Bank corporate finance are gritty incurious sociopathic team players, then that's just who they'll hire. (Here are a couple of sample questions.) The human judgment has been taken out of human resources. (I mean, not entirely; there are still interviews and whatever.) The narrative is gone; the computer just decides what signals predict a good worker, and the workers themselves -- and their bosses -- don't have to understand what the signals mean. Right now the surveys seem to mostly test for characteristics that sound like they'd predict good workers -- grit, etc. -- but there is no reason to think that's the end state. If it turns out that liking electronic dance music, or cheese, predicts success at Deutsche Bank, why wouldn't that go into the algorithm?

Elsewhere in banking personnel, Goldman has reshuffled a lot of its sales and trading leadership. And elsewhere in alienating algorithms, Cathy O'Neil's book "Weapons of Math Destruction" just came out, and is very good -- the unusual book that changed my mind about things. And here is a story about gig-economy workers going on strike called "When your boss is an algorithm":

It’s hard to spread the word when you don’t even know who your colleagues are. But the couriers have an idea. They open their apps as customers and order food to be delivered to them. As UberEats couriers arrive with pizzas at the place their app has sent them, the strikers tell them about the protest and urge them to join in. Algorithmic management, meet algorithmic rebellion.

Fintech: Blockchain.

I have no idea what is going on here but it involves IEX Group, and gold, and the blockchain, which as far as I can tell are numbers 1, 2 and 3 on the list of everyone's favorite financial topics to argue about:

The protagonists of Michael Lewis's book, "Flash Boys: A Wall Street Revolt," are planning a gold exchange that would use elements of blockchain technology to improve transparency and the clearing and settling of trades, said Matt Harris, a managing director at Bain Capital Ventures. Bain has an investment in IEX.

You can more or less divide markets into those that trade abstract entitlements to future cash flows (stocks, bonds, derivatives) and those that trade actual physical objects (gold, cars, blood); one thing that I tend to think is that blockchain is a good system for registering abstract entitlements, but offers few special advantages in moving physical objects. I guess the blockchain-for-tuna people would disagree with me. Of course, if you trade gold, it's not like you (usually) expect to end up with a pile of gold doubloons in your closet. You expect to have an abstract entitlement to a quantity of gold, held for you by someone else. But buying gold tends to correlate with a certain distrust of the normal financial system. An immutable, auditable record of how much gold you own and where it's held would probably be appealing. 

Elsewhere in blockchain, blockchain bootcamps are hot:

With banks and insurers starting to tinker with the blockchain, as a tool to record transactions and asset transfers, and venture capitalists investing more than $1.1 billion in related startups, there aren’t enough developers who have mastered the software. The career site Indeed.com listed 136 jobs with “blockchain” in the description as of Sept. 7, everywhere from New York to Boston, while Monster.com posted 77 jobs.

There's that famous Harvard Business School contrarian indicator -- the sectors where HBS graduates go are the ones that are about to decline -- but has anyone studied the predictive power of coding bootcamps?

Fintech: Robo-advisers.

When markets went crazy after the Brexit vote, robo-adviser Betterment LLC suspended client trading for a few hours, which I think is just a great idea. The most dangerous thing retail investors do is sell after market crashes (hot tip: sell before!), and a popular narrative is that human financial advisers earn their keep during times of crisis, when investors panic and want to sell, and the advisers get on the phone and talk them off the ledge. A robot, the theory goes, won't be as comforting on the phone. But Betterment's approach of not answering the phone -- not letting people trade during crashes -- is even better at preventing panicky retail investors from selling right after a crash. I don't know if that was their intent, or if they would quite put it that way. But I like it.

You know who doesn't? Financial advisers who use Betterment as a platform for their clients:

“It’s a little disheartening,“ said Eric Roberge, a Boston-based fee-only adviser who manages money for some clients through Betterment’s institutional service.

Roberge and others are still waiting for a better explanation from Betterment of its trading-halt policy. Meanwhile:

For now, Mr. Roberge said he plans to limit the number of clients he manages through the platform by, for example, not using it for people who are newer to investing and may require more frequent portfolio changes during a volatile market. “It would be a sticky situation, especially when I don’t know how it works,” he added.

Wait, people who are newer to investing are the ones who "may require more frequent portfolio changes during a volatile market"? Is that ... is that really how retail financial advising works? I am becoming more and more fond of the robots.

Fintech: Put-spread collar index fund!

Last year I joked a bit about Vest, a fintechy startup that seemed to be all in on "selling listed single-stock options to millennials." Vest's co-founder talked down the benefits of diversification, and talked up the benefits of listed single-stock options, which, as a former equity derivatives salesman myself, I found a bit odd.

Anyway Vest is now "a CBOE company" and seems to have embraced diversification, judging by its new CBOE S&P 500 Buffer Protect Strategy Fund, a put-spread collar index fund. And I am telling you about it because I like saying "put-spread collar index fund." It's like an index fund -- so normal! -- but it's also a put-spread collar -- so derivatives-y!

Anyway yeah the point is that you get exposure to the S&P 500 index, but you get rid of the first 10 percent of downside risk. So if the S&P falls by 6 percent, your return is zero. If it falls by 16 percent, your return is negative 6 percent. In exchange, you give up upside above the first 11 percent or so. If the S&P rises by 6 percent, your return is 6 percent. If it rises by 16 percent, your return is 11 percent or so. (It's divided into 12 monthly tranches, and each tranche has its own cap, ranging recently from 11 to 14 percent.)

I like it! How could I not; I am a former derivatives salesman. It is recognizably good derivatives patter: Vest has identified some fears that people have (near-term midsize stock market decline), and has built a payoff curve that addresses them in a plausibly appealing way. Though as a derivatives salesman at some point I kind of went through that way of thinking and came out the other side. If you want exposure to the S&P 500, but you don't want to lose as much money if it goes down, and you are willing to make less money if it goes up, you can just buy less of it. That doesn't give you exactly the same exposure as a put-spread collar index fund, but it is cheaper, and what was it that made you think that was the exact exposure you wanted anyway?

Elsewhere in fintech generally: "You've Heard of Fintech, Get Ready for 'Regtech.'" And: "SoFi Looks to Raise $500 Million in Latest Test for Fintech."

The unbanked.

I suppose that this Wall Street Journal article about how lawyers at big law firms are losing their private-banking status at JPMorgan -- partners are being demoted from "J.P. Morgan Private Bank" to "J.P. Morgan Private Client Direct," while associates are being kicked all the way down to "Chase Private Client" -- will not upset too many people, but the people it does upset will be intensely upset. I sympathize: As a young law firm associate, I was thrilled to receive fancy private-banking status at JPMorgan. (I opened a checking account.) And in some ways the biggest milestones in the last decade or so of my career have been my successive declines in status at JPMorgan. My ATM card now just says "Premier Platinum," which I suspect is many levels below even "Chase Private Client," and just a step or two above "Garbage." I feel your pain, law firm partners who have been cast out of the promised land of private banking.

In happier news: "Ranks of the ‘Unbanked’ Decline, FDIC Survey Finds."

People are worried about non-GAAP accounting.

Good news, the problem is fixed: "On May 17, 2016, the Securities and Exchange Commission issued new Compliance and Disclosure Interpretations ('C&DIs') on the use of non-GAAP financial measures," and a survey of 100 earnings releases by large companies found "a high degree of responsiveness by companies in conforming their disclosure to the C&DIs," with 79 of those companies having "altered the presentation of non-GAAP measures in their earnings releases."

People are worried about unicorns.

"Snapchat has been talking to investment bankers about filing for an IPO towards the end of this year or early in 2017," and if they file a prospectus for an initial public offering and then end up pulling it, I will cry, because of all the jokes about disappearing IPO filings. So good luck with your IPO, Snapchat. Actually can you imagine those pitch meetings? Bankers famously wore yoga pants to pitch for Lululemon's IPO. I hope they show up at Snapchat nude.

People are worried about bond market liquidity.

No, it is silent as the grave on bond market liquidity, but I am going to continue to throw hoo-boy-negative-yields-are-weird-aren't-they articles under this heading. Here is Bloomberg Gadfly's Lisa Abramowicz on the prospect of negative-yielding junk bonds, which sounds even weirder when you remember that the nice way to say "junk bond" is "high-yield bond." How do you like your negative-yielding high-yield bonds? And here is an explainer on why people might buy negative-yielding bonds:

Some investors may speculate that yields can fall even further, which will push up the prices of corporate bonds. So investors in the Henkel and Sanofi bonds may be able to sell them on at an even higher price later.

One reason to think yields could go lower it that the ECB is buying corporate bonds.

Like we talked about yesterday, negative yields really are a liquidity story: If you buy a negative-yielding bond, your only way to make money is if secondary trading markets remain robust. Like, for instance, if there's a price-insensitive source of demand that will buy up any bonds that you want to sell.

Things happen.

European Central Bank Leaves All Three Rates Unchanged. Dell Gets Bigger and Hewlett Packard Gets Smaller in Separate Deals. Billionaire Takes Pole Position in Formula One. Wells Fargo to Be Subject of Enforcement Action Over Cross-Selling, Sales Practices. The Trials and Tribulations of the Vatican’s Finance Chief. BlackRock Says Bond Market Views Puerto Rico Board as Positive. No Man’s Land for Greek Bonds as Tsipras Vacillates on Reforms. This Hedge Fund Made 2,100% From World's Most Extreme Market Mania. Abu Dhabi Investor Mubadala Posts $1.2 Billion First-Half Loss. PIK toggle notes. Hanjin Shipping’s Troubles Leave $14 Billion in Cargo Stranded at Sea. Bad Bad Contagion. Selling a $5 Million, Seven-Story Basket Is No Picnic. Catfished by catfish. Puppy and cheetah. Quit your job. Log flumes.

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

(Corrects title of book, "Weapons of Math Destruction," in 10th paragraph.)

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:
James Greiff at jgreiff@bloomberg.net