Editor's Note: Two more weeks off before I resume faiV duties. This week's guest editor is Alexander Berger, managing director of the Open Philanthropy Project, which "identifies outstanding giving opportunities, makes grants, follows the results, and publishes our findings." And he's a long-time reader and evangelist for the faiV, so you know he's a great guy.-Tim Ogden
1. Universal Basic Income (unpopular locale edition): In 2010, to replace massive energy and food subsidies, the Iranian government apparently implemented a cash transfer program that began covering over 95% of the population (75 million people) before targeting seems to have lowered coverage to less than 35 million. The story in two sentences: “In 2011, the first full year of the program, transfers amounted to 6.5% of the GDP and about 29% of the median household income. After three years of inflation, the amount transferred per person is down to less than 3% of GDP per capita.” New research finds minimal effects on labor supply or hours worked, though the short time horizon for the large transfers makes it hard to generalize. I suspect that the short time horizon is only part of the reason this policy hasn’t gotten more attention.
2. Our AI Overlords: Another AI benchmark falls. In a much-publicized practice event Sunday, an AI system developed by OpenAI beat a team of former pros at a mutliplayer video game called Dota (they had a livestream and posted a video that is totally inscrutable to me). This was expected given the rapidly-growing computation devoted to experiments like this, though it looks like the training required by this model (190 “petaflop/s-days,” whatever those are) was less than would be expected from extrapolating past large experiments. (The costs for those experiments are also growing by an order of magnitude every year and a half, which seems… unsustainable.) Apparently OpenAI are planning a Dota match against current pros later this month, so expect to hear more about this.
3. Cryptocurrency (or Weird Household Finance): Apparently “proxies for investor attention strongly forecast cryptocurrency returns,” which seems… a little obvious? And Matt Levine discovers a subculture of people who intentionally participate in pump and dump schemes in marginal cryptocurrencies as a form of gambling, which raises the question - what do the other people investing in marginal cryptocurrencies think they are doing?
4. Unexpected Results, Incarceration Edition: In Colombia, having their parents incarcerated *increases* educational attainment in kids. In Ohio, having a parent or sibling incarcerated reduces high school graduation rates, along with the probability of both childhood and adult incarceration. In Norway, being sent to prison reduces the probability that a younger brother will be charged with a crime by 32 percentage points. Basically none of these findings are what you’d expect just based on observational data. As usual it’s hard to say what these findings add up to, but there are enough papers using this approach now that it might be time to start asking what we can learn from the whole lot. A question for my colleague David Roodman.
5. Public and Private Markets: On the private market side, Matt Levine had fun this week with Elon Musk’s proposal (threat? joke?) to take Tesla private, perhaps vindicating his view that private markets are the new public markets. On the public market side, this blog post by Jesse Livermore changed how I thought about the equity premium, arguing that it had been formerly justified by the riskiness of individual stocks and the difficulty of indexing and that the low cost and rising share of passive indexing today can help explain (and rationalize) rising valuations and lower expected returns going forward.
In another instance of news from places we usually don't pay attention to, Greece is going through something that, in today's style, should probably be called The Meg Depression. Via Adam Tooze. Source: IMF, and this blog post.
