5 ms·

It's probably worth applying some scepticism to this study: 1. It's an economics paper. Not a field with a reputation for accurate predictions beyond the basic

It's probably worth applying some scepticism to this study: 1. It's an economics paper. Not a field with a reputation for accurate predictions beyond the basics. 2. It's commissioned by a charity that wanted academic evidence of the benefits of its own policies, so it's as conflicted as can be. 3. It's research by academics who are as a group notoriously left wing, so they're being asked to do research to support their own ideological preconceptions. That doesn't mean the conclusions are wrong, but it does mean they shouldn't be accepted at face value, and you'd have to review it very thoroughly - ideally with external double checks of validity - to be able to conclude anything from it. Doing even a very fast check of the paper immediately reveals a major problem: the headline claim of 0.1% price inflation excludes the price of labour, which inflated significantly. This pretty much explains the entire effect. Not surprisingly, if you start handing out cash at random in poor communities then the people who aren't lucky enough to get it start demanding some of that cash too. But you don't need research to know this. It's economics 101, that's what's happening right now in western economies (big cash transfers from governments -> inflation). We may suspect the paper is being undermined by ideology and funder bias because its early statements on inflation try hard to hide this fact. At the start, it's saying: "Importantly, we document ... minimal price inflation" (abstract). Then it becomes: "For inputs, we find positive point estimates [of inflation], but they are not always statistically significant. For outputs, we document statistically significant, but economically minimal, local price inflation." (page 4). Unless read very cynically this sounds like good news. But then we notice how obfuscated this statement is. "Positive [input] point increases but they are not always statistically significant", i.e. some are and the magnitude of the increase isn't mentioned. Later we find: "we do see significant increases in the factors that we directly measure, and particularly in the wage bill: enterprises in treated (control) villages increase spending on labor by USD PPP 82 (70), a sizable change relative to the control mean" (page 20). So we've gone straight from "minimal price inflation" to "significant increases ... a sizeable change". I've read so many academic papers in the last few years and whenever you see such large slippage between claims in the abstract and claims buried deep in the paper, it's always a bad sign. It's a leading indicator that there's going to be other kinds of problem, probably involving bogus stats/models/assumptions. Moving on. We might have hoped that people would use the money for investment that could permanently increase their standards of living after the cash transfers end, but that doesn't happen: "Strikingly, we do not see strong evidence of a firm investment response" But this isn't something that needs research to know, it's the story of decades of massive and sustained foreign aid to Africa. Aid money gets spent on increased consumption but not sustainable investment.

notahacker 4y ago
> So we've gone straight from "minimal price inflation" to "significant increases ... a sizeable change". I'm not sure the fact that a different variable (total wage expenditure) from the one discussed in the abstract (price inflation) recorded a different outcome is a particularly pertinent critique of the paper. Since the paper also records that the increased wage expenditure allowed wage earners to buy more goods because the price of goods inflated minimally, and the enterprises spending more on wages nevertheless experienced a [non-statistically significant] increase in profit, leaving all groups materially better off, this is GDP growth, not inflation. A more pertinent observation is that this is pretty much what you'd expect from injecting additional money into a local economy operating well below capacity and buying most of its consumer goods from areas not receiving cash injections, so it doesn't necessarily work at national scale or if the jurisdiction self-funds its poverty alleviation funding.
- origin_path 4y ago
  Although I agree with your conclusions that the study doesn't tell us much, I don't think wages should be excluded from claims about price inflation, especially in non industrial economies heavily reliant on cheap labour. The charity advertises the headline 0.1% rate and the abstract makes no mention of the fact that actually some of the most important prices in that society increased significantly. If they'd admitted to this up front of would have seriously reduced the apparent success of their initiative. But there is really a deeper problem here. Every time researchers do this they're training people to assume their claims are deceptive in some way. Large groups of people are just tuning out academic claims because of this sort of thing, they don't care. But this is bad for social cohesion because the people who take academic output on faith then conclude that they must be a superior breed of person: "reality based", "understands the science" etc. We already have this problem and it's getting worse. To wit: you can't afford the time to double check every claim presented as important or that will affect social policy that affects you, so you have to generalize, and increasingly that means assuming that if an academic makes a claim convenient for their prevailing ideology, it's probably a trick or misleading in some way. If caught they tend to blame journalists for "misrepresenting" their work, or they'll point to a footnote on page 67 where they redefine a standard term and use it to claim nobody should have ever assumed the obvious interpretation of what they were saying. It's just so tawdry. Then for people who stop listening, it gets used as a weapon to beat them around the head. So I think academics have a moral obligation to be brutally honest in their claims and abstracts. This sort of word game where they arbitrarily exclude the prices that went up from their definition of price inflation, is ultimately self defeating.

notahacker 4y ago

> So we've gone straight from "minimal price inflation" to "significant increases ... a sizeable change". I'm not sure the fact that a different variable (total wage expenditure) from the one discussed in the abstract (price inflation) recorded a different outcome is a particularly pertinent critique of the paper. Since the paper also records that the increased wage expenditure allowed wage earners to buy more goods because the price of goods inflated minimally, and the enterprises spending more on wages nevertheless experienced a [non-statistically significant] increase in profit, leaving all groups materially better off, this is GDP growth, not inflation. A more pertinent observation is that this is pretty much what you'd expect from injecting additional money into a local economy operating well below capacity and buying most of its consumer goods from areas not receiving cash injections, so it doesn't necessarily work at national scale or if the jurisdiction self-funds its poverty alleviation funding.

origin_path 4y ago
Although I agree with your conclusions that the study doesn't tell us much, I don't think wages should be excluded from claims about price inflation, especially in non industrial economies heavily reliant on cheap labour. The charity advertises the headline 0.1% rate and the abstract makes no mention of the fact that actually some of the most important prices in that society increased significantly. If they'd admitted to this up front of would have seriously reduced the apparent success of their initiative. But there is really a deeper problem here. Every time researchers do this they're training people to assume their claims are deceptive in some way. Large groups of people are just tuning out academic claims because of this sort of thing, they don't care. But this is bad for social cohesion because the people who take academic output on faith then conclude that they must be a superior breed of person: "reality based", "understands the science" etc. We already have this problem and it's getting worse. To wit: you can't afford the time to double check every claim presented as important or that will affect social policy that affects you, so you have to generalize, and increasingly that means assuming that if an academic makes a claim convenient for their prevailing ideology, it's probably a trick or misleading in some way. If caught they tend to blame journalists for "misrepresenting" their work, or they'll point to a footnote on page 67 where they redefine a standard term and use it to claim nobody should have ever assumed the obvious interpretation of what they were saying. It's just so tawdry. Then for people who stop listening, it gets used as a weapon to beat them around the head. So I think academics have a moral obligation to be brutally honest in their claims and abstracts. This sort of word game where they arbitrarily exclude the prices that went up from their definition of price inflation, is ultimately self defeating.