Week of February 12, 2018

Editor's Note: I'm obviously not anti-bank (at least I hope that's obvious!), but in the wake of last week's piece on how hard it is to figure out the value of most of what banks do, I've been accumulating a number of pieces on bank behavior that are less than flattering. I've been struggling to come up with any other service that is so vital and that society so commonly holds in contempt. It's a reminder, again, of what an enormous accomplishment it was for microfinance's pioneers to get people to view banks and bankers as heroes. If there are any sociologists in the house who would like to school me on the literature of social perceptions of banking, please do!--Tim Ogden

1. Banking: In case you missed it, here's that link from last week finding that banks would be better off if they did a lot less. Well, a lot less of the complicated financial stuff that most (large) banks spend a lot of time doing. Matt Levine sees a generalized trend in a positive direction--that is that the financial engineering that financial services companies are engaged in is focused much less on engineering complex financial instruments and a lot more on software and technology engineering. Even the cool project names are being reserved for technology projects rather than hard-to-understand derivatives-of-futures-of-insurance-of-bonds-of-weather-derivatives.
That does raise some questions about the evolution of fintech--if the banks themselves are more focused on the technology of service delivery, what does that mean for the technology firms? I do feel a bit of unease that these are the same banks that don't seem to be able to add value to themselves in their core area of expertise (and it's not just the banks, remember that Morningstar's ratings are negative information). How much should we expect from their wading into technology and advice? More on that below, in item 2.
There's another concern with banks moving in this direction. While it's not always the case, the kind of engineering that banks are doing now tend to increase consolidation: returns to scale tend to be bigger and matter more in software, data and high-volume/low-margin activities. And when consolidation happens it tends to be bad for lower-income customers. Here's a recent paper examining the impact of bank consolidation in the US (particularly large banks acquiring small banks): higher minimum account balances and higher fees, particularly in low-income neighborhoods. Those neighborhoods see deposits flow out of bank accounts (justifying closing branches) and later see increases in check-cashing outlets and decreased resilience to financial shocks. But wait there's more: the current version of the Community Reinvestment Act regulations tend to focus on places where banks have a physical presence. So closing branches and delivering more services through technology means, well, that those banks have less worries about complying with CRA. Hey did you know that the Treasury Department is considering making changes to the CRA regulations? I'm guessing the first priority isn't going to be expanding the CRA mandates.
And just to throw in a little non-US spice, here's a story about massive bank fraud at the Punjab National Bank in India.


2. Our Algorithmic Overlords: I've made jabs in the faiV pretty regularly about fintech algorithms ability to make good recommendations, particularly for lower income households. It turns out I'm not alone in distrusting machine-generated recommendations. Human beings tend to believe pretty strongly that humans make better recommendations than machines particularly when it comes to matter of taste. But we're all wrong. Here's a new paper from Kleinberg, Mullainaithan, Shah and Yeomans testing human versus machine recommendations of jokes(!). The machines do much better. Perhaps I should shift my concern away from machine-learning-driven recommendations and spend more time on a different preoccupation: why humans are so bad at making recommendations. There is perhaps another way: making humans and machines both part of the decision-making loop. A great deal of work in machine learning right now is organized around humans "teaching" a machine to make decisions, and then turning the machine loose. An alternative approach is having the "machine-in-the-loop" without ever turning it loose. That is the approach generally being used in such things as bail decisions. The big outstanding question is where we should allow humans (and which humans) to overrule machine recommendations and when we should allow the machines (and which machines) to overrule the humans.
Key to making such decisions is whether the human is able to understand what the machine is doing, and whether humans should trust the machine. Both are dependent on replicability of the AI. You might think sharing data and code in AI research would be standard. But you'd be as wrong as I was about recommendations. There's a budding replication crisis in AI studies because it is so rare for papers to be accompanied by the training data (about 30%) used in machine-learning efforts, much less the source code for their algorithms (only 6%!). Of note if you click on the paper above about recommendations, on page two  there is note that all of the authors' data and code are available for download.

3. Risk: Last week I promised some more thoughts on risk and aspirations. To summarize for those who haven't been following along: there is strong evidence of large returns on investment for poor farmers and even some microenterprises, there are similarly large returns for rural farmers investing in migrating to urban areas, those folks tend to avoid making such investments, and interventions that reduce risk or allow pre-commitment tend to increase such investments. More recently, several other pieces of evidence seem to be falling into place. First, those large returns on investment are not so large once you adjust them for risk (that's from the recent Townsend paper that Jonathan first linked two weeks ago). Second, urban migration might be riskier than we have appreciated. Third, people who migrate may be systematically different and more capable (and thus have less risk) than those who don't. And fourth, as I talked about last week, another way to get people to make more investment is to raise their aspirations or sense of personal efficacy, which could be interpreted as increasing their risk tolerance. 
There are a number of things that strike me given this set of stylized (and not yet fully proven) facts:
1) There are big reasons to be concerned about general equilibrium effects of increasing the risk tolerance of people who are risk averse. It's very plausible that early experiments in this domain would show large gains for participants but those gains would not only fade, but substantially reverse at scale. If this pattern is true, it makes a very, very strong case for investing in insurance that protects people from risk rather than changing their risk tolerance.
2) The pattern of risk-adjusted returns in the Townsend data looks a lot like the entrepreneurial equilibrium in developed countries, as described by Amar Bhide in his book The Origin and Evolution of New Businesses. The short version: established businesses take all of the less risky investments, leaving the truly high risk ones to entrepreneurs. Those entrepreneurs take them not only because they are less risk averse, but because they are the only options available--which is consistent with general findings that successful entrepreneurs are just as risk averse as corporate managers. But we only ever see the risks which pay off, leaving us with a profoundly distorted view of entrepreneurism.
3) A book I spent some time reading while on break was James Scott's Against the Grain. One of the main claims of the book is that, contrary to the traditional narrative, the hunter-gatherer lifestyle is far less risky than the sedentary agricultural lifestyle. He makes a very convincing case using all sorts of evidence, but it raises the big question of why risk-averse agriculturalists haven't continually reverted back to hunter-gatherer lifestyles. I find the arguments there less convincing, but I'm not sure what to think about that or the implications. 

4. Inequality: OK, late in the day. Time for some rapid fire links. An argument for income redistribution to address growing inequality from an unexpected source: Bain & Company. A new paper looks at whether one of the methods for income redistribution, a Universal Basic Income, discourages work by examining Alaska's citizen oil dividend and finds that it mostly doesn't, though with some effects in tradeable sectors. Overlooked in many discussions of inequality is the largest disparity in college-going: rural kids are the ones most left behind.
And a lengthy piece on the hidden inequality in how people in the US make payments. The rich use credit cards and get lots of rewards (like cash back or airline miles, mostly paid for by merchants, and don't carry balances) and the poor use debit cards or cash and get nothing--making for a very regressive system. Just another way the poor are different than you and me: they pay more.

5. Microfinance Is Just Banking: And to tie the whole thing together, I'm going to close with two pieces about microfinance. First on the motivations and impact of an MFI in Bangladesh dropping its group meetings in favor of mobile money transactions. Second on what's wrong in Sri Lanka's microfinance industry. Is it a Straussian reading if I tell you to read this item like a Straussian? Here let me make the sub-text text: you should read these pieces only after looking at the pieces above what is wrong with banking and be amazed at how quickly microfinance seems to be re-learning all the lessons of modern consumer banking that are evident in developed countries or even in other countries with more mature microfinance.

  Mechanical Turk is a common source of 'warm bodies' for social scientists but it's hard to know to just who the workers who participate are, and even how many there really are--and the answers are highly dependent on how you define who is a "regular" participant. It's a complicated question to answer, and the chart below is interesting but wrong,  for interesting reasons . Source:  Panos Ipeirotis .

Mechanical Turk is a common source of 'warm bodies' for social scientists but it's hard to know to just who the workers who participate are, and even how many there really are--and the answers are highly dependent on how you define who is a "regular" participant. It's a complicated question to answer, and the chart below is interesting but wrong, for interesting reasons. Source: Panos Ipeirotis.

Return to the Weekly faiV