Editor's Note: I had been trying to get this out early given the holiday weekend in many places, but so it goes. One of the struggles of getting out the faiV is how to connect items together in something resembling a narrative flow. For this edition, it was especially difficult as there are two “diptychs” that are related, but not wholly overlapping—and I got myself turned into knots trying to figure out how to write about them. I haven't solved the problem so much as given up on it. So forgive me for this sub-optimal presentation.
If you're ready to for a happy start to your week, scroll to the end. No April Fools here. - Tim Ogden
1. Small Firm Diaries
This week we published the Small Firm Diaries Global Analysis slide deck. We're getting this out well ahead of a formal (and long) report, in part as we try to adjust to how people consume and use research now. The deck covers the four key findings, and related recommendations, with enough detail that you can follow the story, we hope. And we hope that putting it in this format makes it easier for people to both consume and to use the charts and ideas in their own work. Please do let me know what you think.
We also hosted a faiVLive this week to go over those findings and discuss “what's next” in the world of small firms with Paddy Carter, Payal Dalal, Muhammad Meki and Hillary Miller-Wise. You can view the recording here for the next week and on our YouTube channel thereafter.
We had a great conversation during the faiVLive and got some really interesting questions and comments during and after. We're writing those up and will be sharing them with subscribers to the Small Firm Diaries update list, so it you're not one of those people, go ahead and take care of that now. If you did attend, and are looking for more on these issues, check out the the new podcast from the Argidius Foundation (one of the funders of the Small Firm Diaries) on access to finance for small firms.
2. Payments and FinTech, Part I
We'll start the first diptych with a specific instance of a particular issue in the payments and fintech world—dynamic pricing. This hasn't attracted much attention in the development world's focus on mobile money and digitizing payments—where we've talked much more about the transaction efficiency, security and informational gains—but it is absolutely a feature of digital payment systems: they enable the ability to change prices much more rapidly. Most people are now familiar with dynamic pricing from "surge pricing" of ride share and delivery; you may also have heard about the uproar over Wendy's announcement about trialing dynamic pricing. But it also applies to such things are the pricing of mobile money transactions.
IPA has a post about Airtel in Uganda raising its fees in a particularly unfortunate way: little to no customer notification and little to no transparency. Without digitization, you couldn't pull something like this off. With digitization, Airtel can (at least technically) change the price of transactions on a whim. But as IPA notes, this kind of behavior can have dramatically negative consequences on consumer trust, and ultimately use and behavior.
You might think that this is a fairly narrow example. You might believe that Airtel Uganda is making a mistake and there will be consequences in the market. We may very well learn a great deal about this in a lot of places shortly though.
A huge change to the global digital payments infrastructure was announced last week—and I continue to be quite surprised that it's not getting more attention. Visa and Mastercard built their payment systems on some simple rules that made adoption very simple for consumers: merchants could not charge differentially for use of a card over cash, and merchants couldn't charge differentially based on which card a consumer was using (e.g. no surcharge for a card that offers airline miles or other rewards over one that doesn't). That matters because the price a merchant pays per transaction varies based on what bank issues the card and the features of the card.
Visa and Mastercard have now agreed to drop those rules, and merchants will be able to add card surcharges, even differential surcharges based on the issuer (issuers are the banks that you have an account with, not Visa or Mastercard) or based on the cost to them (e.g. more for an airline rewards card).
What's going to happen? I honestly have no idea. There could be “chaos” where consumers have to ask what the cost of using different cards is. Some merchants might offer a discount if you use a card issued by the bank they use as their processor. Some large merchants may choose to ignore the flexibility and keep things as they are, while smaller merchants have differential pricing. Consumers could switch back to cash, or pin-based debit card transactions, to get price uncertainty. Airlines could be left scrambling as revenue from miles cards make up a meaningful part of their profits these days (American remade their loyalty program just last year to emphasize spending on credit cards more than miles flown). Or maybe nothing will happen.
Regardless we're about to learn a lot about payments, fintech and consumer behavior. And about small merchants and cash flow.
3. Payments and Fintech, Part II
To open Part II of this diptych, here's a piece about “the future” of finance, and how fintech is going to drive “hyperefficiency.” Certainly if you've spent any time in the broad financial inclusion space, either domestically or internationally, the lack of efficiency in large parts of the financial services marketplace is a known frustration. But if you've spent any time in this space, there's a different known frustration: fintech hasn't actually done much to remove many of those efficiencies.
It's a long piece and it covers a lot of ground, but there is a common, but wrong, theme: technology can "fix" the inefficiencies. The fact that “technology won't save us” is one of the lessons of my recently published paper (with Joyce Klein of Aspen BOI) on the lessons global microfinance can learn from the US. But more specifically, I was immediately reminded of a piece I linked in the revived faiV a few weeks about “Why Fintech Failed” (and it not coincidentally has a section about payment platforms relevant to the first part of the diptych). For today's purposes you can mostly skip to the end to read, with all the humility that the first piece lacks, that the "fundamental challenges in this space are so strong [that progress will be] without drastic or revolutionary benefits..."
That's really, really not a bad thing. Drastic or revolutionary change in financial services systems should scare everyone. For instance, one of the hopes of digitization—and something that we discussed during the faiVLive—is the possibility that digitization of payments, records and other data can help solve the liquidity challenges of small firms. But use of data is tricky. FinRegLab has a new report on the use of alternative data for small firm lending [they call it “MSE” but if you listen(ed) to the faiVLive you'll hear that I'm on a campaign to permanently bury that term]. One of the conclusions is that a comprehensive alternative data system would be much more likely to reduce exclusion than piecemeal efforts. Of course, getting everyone—banks, processors, firms, regulators, etc.—to agree on such a platform is, shall we say, difficult. Or consider this piece about the boundless possibilities of AI setting insurance premiums—which to it's credit, it notes could be a solution to a lot of insurance market challenges and a dystopian nightmare at the same time.
If you're not worried about the downsides and dangers of a what a comprehensive system of alternative data for underwriting credit and issuing insurance might do, even absent our AI overlords, you obviously have never seen the movie Brazil. Or perhaps you're unaware that markets like this to shut competitors out of online platforms already exist.
Proper caution means moving slowly, but moving slowly also means that hyper-efficiency is not going to happen. And we should probably not be aiming for hyper-efficiency. Inefficiency in some systems is a feature and not a bug (Brazil pun fully intended).
4. Effective Altruism
Full disclosure: I'm chair of the board of GiveWell, so I'm far from objective on this topic. I might have skipped this item entirely if it wasn't a) highly overlapping with a lot of things I cover in the faiV about aid effectiveness, RCTs in economic development, etc., and b) part of the second diptych.
We'll start with an interview of GiveWell co-founder and CEO Elie Hassenfeld by Virginia Heffernan. If you're familiar with GiveWell there won't be much new here, but if you're only vaguely aware and wondering about the approach in the wake of the black eyes that effective altruism got from the OpenAI board revolt and the downfall of Sam Bankman-Fried, this will be a useful overview.
For some reason, Wired decided to pair this interview with an essay by Stanford philosophy professor Leif Wenar, titled “The Deaths of Effective Altruism.” As noted, I'm far from objective, but I was shocked at what a poor piece this is. I'm entertaining the idea that it is itself an experiment in people's reactions to bad arguments; or perhaps an extended joke about Chidi from The Good Place. Long time observers of aid effectiveness debates will have trouble scrapping their jaws off the floor as they read that the author learned of the problems of aid effectiveness while talking to some aid workers while on vacation in Bali. The rest of the evidence mustered doesn't get much, if at all, beyond that level. While arguing that effective altruism doesn't spend enough time considering harms of interventions, for instance, he regularly links to GiveWell's analysis of those harms.
Anyway, it's pretty maddening to me at least, particularly because it essentially justifies not being careful and deliberate (and skepticism of technology-driven hyperefficiency), but inaction.
5. Evidence-Based Decision Making
For those interested in being careful and deliberate in making decisions and evaluating evidence, there are a few things to turn to—after all, academics can't match experienced forecasters, even with prompts, at predicting long-run causal effects.
You could try to be super-rigorous in your evaluation of evidence before drawing conclusions. For instance, by creating an entire platform to systematize Bayesian reasoning. And then you could use that system to draw some conclusions, and then bet people $100,000 that they can't beat you in a debate by convincing impartial judges that your conclusion is wrong. You might be thinking, “that's all taking it a bit too far, don't you think?” And, of course, that is your clue that in fact someone has done this. Here is a very, very, very long write-up about Rootclaim, Saar Wilf, and an epic debate about the origins of COVID.
My favorite part of the whole thing is that on a forecasting website there were bets on whether Wilf would be a “sore loser” if he lost the debate. And those bets paid off. Amazingly, after noting that the article itself is very very long, the comments thread on it is roughly 10x longer. It's things like this that create a receptive audience for Wenar's critique of effective altruists— “please, some one tell me why I don't have to listen to all of this!”
I have a much better approach for you than throwing the baby out with the bathwater.
Jonathan and colleagues have a brand new paper (so new, it's not even on the FAI website yet!) about a careful and deliberative process to take evidence from experimental interventions, and decide where to test them next for maximum usefulness and setting policy. Even better, they are using the framework to choose sites for a replication of an experiment encouraging mobile money-based urban-to-rural remittances.
It is possible to have a grounded, principled approach to learning about the world and implementing policy with due deliberation to make the world better.
Video of the Day
Long-time readers of the faiV know that my older son, Nathanael, has an ultra-rare genetic disease. I remember clearly September 9th, 2005—it was the day, after reviewing an ultrasound, that doctors told us he likely wouldn't survive to be born. This past week had several dates to remember: Nathanael was accepted to Haverford and Davidson, and then he went out and did this (in case you can't read the bib, it says "Blind Skier", and yes, he's completely blind). So start your week remembering that amazing things happen, and there are amazing people in the world at places like the Breckenridge Outdoor Education Center who help make amazing things happen.
Some good vibes for the week.
The faiV is written by Timothy Ogden and produced by the Financial Access Initiative at NYU's Wagner Graduate School of Public Service
Email: fai-wagner@nyu.edu
To read this in your inbox, subscribe to the faiV.
