Week of May 21, 2018

1. Banking: Coinbase, a cryptocurrency trading platform, is doing something strange: acting a lot like a traditional bank by emphasizing its stability and trustworthiness. As Matt Levine points out (save that link, it's going to come up again later),  this is the central paradox of cryptocurrencies--they supposedly do away with the need for trust, but most everyone needs a trusted intermediary to keep hold of their cryptocurrency and protect them against fraud. Y'know the sort of things that banks or governments do (or enable and enforce with regulation). You've probably heard the mantra that cryptocurrencies aren't that important but blockchain is. The Coinbase approach, which is apparently successful, puts the lie to that notion. Why do you need an expensive and inefficient distributed ledger when you can have a cheap and efficient one provided by a trusted intermediary, like Coinbase? 
Trusted intermediaries are really important and the reason why it's worth caring about financial sector deepening. Rather than being distracted by cryptocurrencies, and their inevitable march toward realizing the need for trusted intermediaries, a more fruitful line of thinking is paying attention to what trusted intermediaries are emerging and how they affect consumers, transactions and the flow of money. This was a big part of the story of MFIs success, and one which I think remains underappreciated. Telecoms providing mobile money platforms is a really interesting case, of course. So are the commerce platforms that are rapidly becoming (or already are) payment platforms: Amazon, Google, Facebook, Tencent and Alibaba. And so stories like this about Amazon and this Planet Money story about Tencent and Alibaba (it's called "A Series of Mysterious Packages," how can you resist?) may not seem like they are about banking, but they are about banking.
Beyond the obvious, the reason that the emergence of non-bank but sort-of-like-a-bank trusted intermediaries is that they change the structure of the market. Here's a new paper from de Quidt, Fetzer and Ghatak on market structure and borrower welfare in microfinance, arguing that competition can yield borrower outcomes that match non-profit lending. I'm not yet convinced. And yet, the NY Times Upshot new "Marx Ratio" determines that banks are socialist collectives (that's the Matt Levine link again, I really really wish I could link to specific parts of his posts).
Speaking of market structure, here's a story about American Samoa creating the first public bank in the United States since the turn of last century. Why? I suppose you could say the lack of competition was hurting borrower welfare.

2. Digital Finance: Here's a paper on how using social pressure to encourage positive health behaviors that every MFI that uses groups in any way should read, whether they are doing anything digital or not. There's a U-shape to the curve: the most influential people in changing behavior are those that are neither too close nor too distant in the social graph.
MicroSave has a new piece that gets helpfully specific on the opportunities for using digital finance to close the inclusion gap in six Asian countries (Bangladesh, China, Malaysia, Myanmar, Nepal and Vietnam). There are also country specific reports for most of the countries. Here's something similar from the IFC on Africa. And here's something similar from IIF with a focus on data rather than delivery.
Here is Felix Salmon's interview of the last remaining founder of Simple, one of the first digital banks in the United States, as he prepares to exit. It's mostly a discussion of why it's so hard to be a good bank, while complying with regulations designed to ensure that banks remain trusted intermediaries. Here's a recent announcement from Simple about their Emergency Savings tool which promises to help people figure out the right amount of emergency savings. I'm really curious about how they are really doing that, but the company hasn't responded to my questions.  

3. Our Algorithmic Overlords: Boy there's a lot of stuff built up here so I'm just going to list it out. Here's Erik Brynjolfsson interviewing Danny Kahnemann about algorithmic decision-making and AI. I bet you can guess whether Kahnemann is more worried about bias among humans or algorithms. Here's an article about Judea Pearl's (an important contributor to the development of AI systems) new book, which criticizes the current state of AI development for reasons that will make economists everywhere stand and cheer: AI systems don't understand cause and effect. Of course, per Kahnemann, that's a criticism that could probably be leveled more devastatingly at humans than at machines. Here's Gary Marcus and Ernest Davis writing about how underwhelmed we should be about Google Duplex and that AI is harder than you think. But again, it's plausible to say that AI is harder than you think because human beings are so bad at thinking.

4. Storytelling: Time for a little--I promise--rant from me. Here's a useful piece from SSIR on how to tell stories about complex issues. It's important to tell stories about complex issues and the piece has some good tips. But it includes this line, "no one has ever taken action because of a great graph or data point" which is so demonstrably false that it makes me want to bang my head against a wall until it's bloody. Millions of people--researchers, economists, analysts, engineers, medical professionals, even manufacturing line workers--are taking action every day based on a graph or data point. The statement--the whole paragraph really which suggests that it's common to talk about complex issues with data rather than stories--illustrates exactly what's wrong with storytelling. Stories help us suspend disbelief and critical thinking, and give rise to all the biases that Kahnemann is worried about.

5. Data: So let's talk about some data, eh? Here's Mary Kay Gugerty and Dean Karlan on collecting and using data rather than stories. The new Survey of Household Economic Decision Making from the Fed is out and is required reading for anyone who cares about the state of household finances in the United States. A "good" news headline that got some attention: 40% of Americans can't cover a $400 emergency expense. If you're wondering why that's good news, in 2013 it was 50%. I really can't imagine it's actually true, but I have to wonder if the widespread media coverage of the finding has led to a lot of financial counselors and households setting $400 as the emergency savings target. If you're interested in the data that Jonathan found most interesting, click here, here, here and here.  
As I mentioned a few weeks ago, the new Global Findex is also out. Here's Kaushik Basu on the value of Findex data. If you see other pieces putting Findex data to good use, be sure to let me know.

Via Arianna Legovini, [ready for some Drunk World Bank-worthy convolutions?] the World Bank's Research Group's DIME group's i2i (which is a trust fund of monies from several European aid agencies) has a report on it's work, "Science for Impact: Better Evidence for Better Decisions." I think the graphic is very cool, and interestingly, aligns quite well with short-term poverty spells in the US. I wonder if anyone will take action based on it? There are lots of stories in the report too, just in case. Source: World Bank