Posts

A Local Community Course That Raises Mental Wellbeing and Pro-Sociality 2020-01-31T01:42:27.188Z · score: 16 (7 votes)
A corrected model suggests climate change interventions may be within a factor of two of direct cash transfers 2019-11-26T00:19:29.996Z · score: 32 (15 votes)
The Economic Lives of the Poor 2019-11-20T21:17:52.523Z · score: 32 (17 votes)
The (un)reliability of moral judgments: A survey and systematic(ish) review 2019-11-01T02:38:48.809Z · score: 11 (7 votes)
Model-free and model-based cognition in deontological and consequentialist reasoning 2019-09-23T20:03:48.793Z · score: 8 (4 votes)
The cost of slow growth chickens 2019-09-12T17:46:14.376Z · score: 17 (10 votes)
Uncertainty and sensitivity analyses of GiveWell's cost-effectiveness analyses 2019-08-31T23:33:58.276Z · score: 67 (26 votes)
Consumer preferences for labgrown and plant-based meat 2019-08-08T18:45:05.581Z · score: 18 (7 votes)
Realizing the Mass Public Benefit of Evidence-Based Psychological Therapies: The IAPT Program 2019-07-16T00:25:07.010Z · score: 8 (5 votes)
"Moral Bias and Corrective Practices" and the possibility of an ongoing moral catastrophe 2019-06-24T22:38:15.036Z · score: 12 (4 votes)
Summary of Cartwright and Hardie's "Evidence-based Policy: A practical guide to doing it better" 2019-06-17T21:25:25.006Z · score: 17 (9 votes)
Doning with the devil 2018-06-15T15:51:06.030Z · score: 3 (3 votes)

Comment by cole_haus on Take care with notation for uncertain quantities · 2020-09-16T19:54:36.372Z · score: 8 (5 votes) · EA · GW

Note that the significant figures conventions are a common way of communicating the precision in a number. e.g. indicates more precision than .

Comment by cole_haus on What are examples of EA work being reviewed by non-EA researchers? · 2020-03-24T18:24:42.545Z · score: 14 (10 votes) · EA · GW

In addition to Will MacAskill's critique of functional decision theory (MIRI-originated and intended to be relevant for AI alignment), there's this write-up by someone that refereed FDT's submission to a philosophy journal:

My recommendation was to accept resubmission with major revisions, but since the article had already undergone a previous round of revisions and still had serious problems, the editors (understandably) decided to reject it. I normally don't publish my referee reports, but this time I'll make an exception because the authors are well-known figures from outside academia, and I want to explain why their account has a hard time gaining traction in academic philosophy.

Comment by cole_haus on What are examples of EA work being reviewed by non-EA researchers? · 2020-03-24T18:20:51.342Z · score: 22 (12 votes) · EA · GW

Here's a thread in which a World Bank economist critiques GiveWell on research/publication methods. (GiveWell responds here.)

Comment by cole_haus on AMA: "The Oxford Handbook of Social Movements" · 2020-03-23T03:47:33.593Z · score: 3 (2 votes) · EA · GW

I just feel like it's hard to come away with much of long-term value. I sort of nod along as I read thinking, "That's plausible," and that's about it. (To be concrete: I make Anki cards for most nonfiction I read and I've only made around 1o or 12 across 200 pages which is way fewer than normal for me.) I think I generally want my non-fiction to have at least one of:

1. Solid empirical findings (i.e. widely and repeatedly attested within the field)
2. Falsifiable models with some explanatory depth (i.e. not just mindless curve fitting or a listing of all possible causal factors)
3. Insightful conceptual analysis (e.g. mutually exclusive and collectively exhaustive taxonomies)

Regarding 1, several empirical studies are mentioned but they don't seem to add up to a coherent or even non-contradictory whole.

There's basically none of 2.

The book is probably closest to achieving number 3, but still not great. I would have liked, for example, if they talked about why the classic agenda of "collective action frames", "mobilizing structures", and "political opportunities" is a better organizational scheme than the alternatives.

The book also focuses more on apportioning credit and on the history of the thinking in the field than I'd prefer.

All that said, I understand different readers are looking for different things.

Comment by cole_haus on Lant Pritchett's "smell test": is your impact evaluation asking questions that matter? · 2020-03-19T06:58:34.751Z · score: 2 (6 votes) · EA · GW

I remain pretty confused by this line of argument. I basically parse it as: we should strive to make the actions of developing countries similar to the (best) actions of developed countries. But actions seem of merely instrumental interest and what we actually care about is states (conditions) that are conducive to development.

The recommendations from these two perspectives (actions vs states) converge only insofar as the best actions are invariant across states. But this is quite a big claim and contradicted by e.g. Rodrik who insists that "Institutional innovations do not travel well".

It seems like the development interventions we commonly see can be readily justified by the state-based view. For example, no, we didn't see widespread deployment of insecticidal nets in the US, but, yes, we did see deliberate effort to achieve and good returns from achieving a low burden of infectious disease in the US. No, we didn't have women's self-help groups, but, yes, we did achieve a state of increased gender equality and of increased integration of women into the formal economy.

TL;DR: Why would we expect the same actions to produce the same end state given different starting states?

Comment by cole_haus on AMA: "The Oxford Handbook of Social Movements" · 2020-03-18T20:22:52.323Z · score: 6 (4 votes) · EA · GW

Another book in this area is Handbook of Social Movements Across Disciplines. Unfortunately, I'm most of the way through and it's a bit underwhelming.

Comment by cole_haus on Against anti-natalism; or: why climate change should not be a significant factor in your decision to have children · 2020-02-26T20:14:46.240Z · score: 5 (4 votes) · EA · GW

Here's a half-baked argument for natalism vis-à-vis climate change:

Carbon emissions in the highly developed countries most EAs live in are generally trending in the right direction (i.e. there seems to be at least relative decoupling between emissions and consumption). The bulk of emissions growth over the next several decades will be in other large, rapidly developing countries like India and China. Green technology transfer is a way that highly developed countries can positively influence emissions in the critical rapidly developing countries (see e.g. this). Economic models generally propose that a larger population generates more ideas and a higher rate of technological change (e.g. Population Growth and Technological Change: One Million B.C. to 1990). Therefore, the (smallish?) direct impact of increased emissions from greater population in highly developed countries might be outweighed by more green technology and technology transfer to the crucial rapidly developing countries like China and India.

Comment by cole_haus on Harsanyi's simple “proof” of utilitarianism · 2020-02-20T16:39:35.037Z · score: 37 (17 votes) · EA · GW

Thanks for writing this up!

Comment by cole_haus on Chloramphenicol as intervention in heart attacks · 2020-02-19T01:28:18.846Z · score: 4 (3 votes) · EA · GW

Chloramphenicol is an approved drug, but not approved for this purpose. Approving Chloramphenicol as a coronary treatment requires human trials that will probably cost 25 million. I am extremely far from an expert here so there may be some subtlety, but off-label uses are generally possible. From Wikipedia: However, once a drug has been approved for sale for one purpose, physicians are free to prescribe it for any other purpose that in their professional judgment is both safe and effective, and are not limited to official, FDA-approved indications. This off-label prescribing is most commonly done with older, generic medications that have found new uses but have not had the formal (and often costly) applications and studies required by the FDA to formally approve the drug for these new indications. Edit: The full post at the link acknowledges this: As an approved drug (though for another purpose) any doctor can prescribe Chloramphenicol for any purpose. Of course, they don’t know to do this. And — perhaps more importantly — such bold action can get American doctors sued for malpractice. Comment by cole_haus on Thoughts on electoral reform · 2020-02-19T01:22:58.672Z · score: 34 (11 votes) · EA · GW I just finished reading Democracy for Realists recently which argues that: They demonstrate that voters—even those who are well informed and politically engaged—mostly choose parties and candidates on the basis of social identities and partisan loyalties, not political issues. They also show that voters adjust their policy views and even their perceptions of basic matters of fact to match those loyalties. When parties are roughly evenly matched, elections often turn on irrelevant or misleading considerations such as economic spurts or downturns beyond the incumbents’ control; the outcomes are essentially random. [...] Achen and Bartels argue that democratic theory needs to be founded on identity groups and political parties, not on the preferences of individual voters. I'm not fully settled on how much weight to give to this perspective, but I think it's important to remember the empirical facts of voting as it happens in the real world and not just the idealizations of social choice theory. Presumably this leads to a quite different notion of the optimal electoral system and the optimal series of electoral reforms. (This isn't meant to say that the social choice theory perspective and the points brought up in this post are unimportant. I just thought it was an interesting book and a good reminder to look at this whole other set of criteria.) Comment by cole_haus on If you (mostly) believe in worms, what should you think about WASH? · 2020-02-19T01:03:35.575Z · score: 6 (4 votes) · EA · GW A few of the summary points in Safe Drinking Water for Low-Income Regions are interesting and may provide reason for a bit of pessimism: 1. Safe drinking water from “source to sip” consists of a series of interactions between technologies, their delivery models, their scales and costs of production, and consumer uptake and consistent use. Safe drinking water is a system, not a product or an intervention. 1. It seems unlikely that household treatment and safe storage systems—with the possible exception of boiling—can be transformative at scale under current prices, delivery models, and preferences, but they are effective and protective in specific contexts. 1. Cost analyses for “low-cost” systems are usually reported on a partial basis, with installation costs and some operational costs included. The enabling costs of social marketing, mobilization, education, reminders, and community- or household-based unpaid labor are mentioned but not explicitly accounted for. 1. Delivery models and business models significantly affect costs and uptake, at all three scales of service. Yet they are rarely made explicit. 1. Safe water systems can be highly effective, but consumers undervalue drinking water quality and have low willingness and/or ability to pay for safety. This is a particular challenge for arsenic mitigation or avoidance, as arsenicosis is only evident after several years of exposure. Without an in-depth of understanding of either deworming or WASH, it seems like the implementation issues for WASH may be more severe? Comment by cole_haus on If you (mostly) believe in worms, what should you think about WASH? · 2020-02-19T00:59:32.383Z · score: 2 (2 votes) · EA · GW This doesn't directly engage with the point of this post, but Safe Drinking Water for Low-Income Regions is a pretty good introduction to WASH IMO: Well into the 21st century, safe and affordable drinking water remains an unmet human need. At least 1.8 billion people are potentially exposed to microbial contamination, and close to 140 million people are potentially exposed to unsafe levels of arsenic. Many new technologies, water quality assessments, health impact assessments, cost studies, and user preference studies have emerged in the past 20 years to further the laudable goal of safe drinking water for all. This article reviews (a) the current literature on safe water approaches with respect to their effectiveness in improving water quality and protectiveness in improving human health, (b) new work on the uptake and use of safe water systems among low-income consumers, (c) new research on the cash and labor costs of safe water systems, and (d) research on user preferences and valuations for safe water. Our main recommendation is that safe water from “source to sip” should be seen as a system; this entire system, rather than a discrete intervention, should be the object of analysis for technical, economic, and health assessments. Comment by cole_haus on What can the principal-agent literature tell us about AI risk? · 2020-02-10T22:51:02.122Z · score: 2 (2 votes) · EA · GW This isn't looking at it from exactly the same angle as this post, but Incomplete Contracting and AI Alignment also looks at the alignment problem through the principal-agent lens: We suggest that the analysis of incomplete contracting developed by law and economics researchers can provide a useful framework for understanding the AI alignment problem and help to generate a systematic approach to finding solutions. We first provide an overview of the incomplete contracting literature and explore parallels between this work and the problem of AI alignment. As we emphasize, misalignment between principal and agent is a core focus of economic analysis. We highlight some technical results from the economics literature on incomplete contracts that may provide insights for AI alignment researchers. Our core contribution, however, is to bring to bear an insight that economists have been urged to absorb from legal scholars and other behavioral scientists: the fact that human contracting is supported by substantial amounts of external structure, such as generally available institutions (culture, law) that can supply implied terms to fill the gaps in incomplete contracts. We propose a research agenda for AI alignment work that focuses on the problem of how to build AI that can replicate the human cognitive processes that connect individual incomplete contracts with this supporting external structure. Comment by cole_haus on Clean cookstoves may be competitive with GiveWell-recommended charities · 2020-02-10T22:46:27.868Z · score: 16 (7 votes) · EA · GW I didn't see it among your links, but GiveWell has an interim intervention report on this. Their summary is: • What is its evidence of effectiveness? Results from multiple randomized controlled trials (RCTs) suggest that distributions of clean cookstoves do not have clear evidence of effectiveness at reducing health problems attributable to air pollution. The evidence we have reviewed in our preliminary investigation finds limited impacts on women’s health and no clear impacts on children’s health under typical use. Distributions of clean cookstoves may have been less effective than expected due to implementation challenges, such as low compliance with using the replacement stoves and failure of the cleaner stoves to reduce air pollution sufficiently. • How cost-effective is it? We have not produced a cost-effectiveness model for clean cookstoves because we have not yet seen strong enough evidence to model a health benefit of the intervention. Comment by cole_haus on A Local Community Course That Raises Mental Wellbeing and Pro-Sociality · 2020-01-31T22:42:29.371Z · score: 2 (2 votes) · EA · GW I'm also pretty uninformed on the biomarkers aspect. Beyond what they say in the paper: One reason why we do not find significant effects on biomarkers at conventional levels may be power issues combined with relatively noisy measures. Another, related reason may be the composition of our sample: high levels of pro-inflammatory cytokines have been found for major depression; respondents in our sample, however, report, on average, only mild depressive symptomatology, pre-treatment. In fact, we find that only eight out of 133 respondents (about 6%) report strong depressive symptomatology, as indicated by PHQ-9 scores of fifteen or higher. Moreover, even amongst these, only about a third show associated elevated inflammation (Wium-Andersen and Nielsen, 2013). For cortisol, individual differences and timing of measurement matter; it has been found to be a rather short-term measure for stress (Miller et al., 2007). I found Table 5 here gives a bit more context on the correlation between these biomarkers and different outcomes. Comment by cole_haus on A Local Community Course That Raises Mental Wellbeing and Pro-Sociality · 2020-01-31T22:26:54.957Z · score: 2 (2 votes) · EA · GW trained facilitator This is how they describe their facilitators: The course is manualised and scalable: each course is led by two volunteers – screened by Action for Happiness for motivation and skills, and once approved, provided with structured resources – as facilitators on an unpaid basis in their local communities. Recruitment of course leaders follows a carefully documented, standardised process: each candidate completes a Leader Registration process sharing their motivation and skills and is given clear instructions on what is required. Once potential course leaders have a co-leader, venue, and dates in mind, they complete a Course Application process. The team at Action for Happiness reviews this application and, if all criteria are met, arranges a call to discuss next steps. Once a course is fully approved, course leaders receive on-going guidance and support. There is also a post- course follow-up process. Not sure if that's what you had understood and meant with 'trained facilitator' (just wanted to make it clear that it doesn't mean licensed behavioral therapist or something). Comment by cole_haus on A Local Community Course That Raises Mental Wellbeing and Pro-Sociality · 2020-01-31T22:15:32.334Z · score: 2 (2 votes) · EA · GW As far as comparisons, they say: Impacts on subjective wellbeing, mental health, and pro-sociality are large: the course increases life satisfaction on a zero-to-ten scale by about one point, more than being partnered as opposed to being single (+0.6) or being employed as opposed to being unemployed (+0.7) (Clark et al., 2018). It is more than double the effect of ENHANCE, a 12-week course focusing primarily on positive habits, skills, and attitudes, which is probably the most comparable intervention (Kushlev et al., 2017). 28 However, the authors are able to track outcomes over a longer period of time, up to six months post-treatment. Finally, the effect on life satisfaction is somewhat larger than effects found in trials by the UK Big Lottery Fund, which funded a wide range of wellbeing programmes (fourteen portfolios, each consisting of three to 34 actual trials) from 2008 to 2015 at a volume of £200 million. Trials typically included community-based activities such as horticultural activities, cooking lessons, or sports events. As a conservative estimate, they increased life satisfaction on a zero-to-ten scale by, on average, 0.5 points for six months post-treatment (New Economics Foundation-Centre for Local Economic Strategies, 2013). Different from our intervention, however, these trials all targeted specific groups with mental health needs, including overweight adults, families with young children, or people with substance use disorders. Comment by cole_haus on A Local Community Course That Raises Mental Wellbeing and Pro-Sociality · 2020-01-31T22:11:59.938Z · score: 2 (2 votes) · EA · GW Thanks for your thoughts! Yes, regarding persistence they also note: To look at treatment effect persistence, we exploit data points at follow-up in an extended sample. As all respondents have been treated at follow-up, we cannot estimate causal effects, so that results are exploratory. Comment by cole_haus on Oddly, Britain has never been happier · 2020-01-30T23:24:14.417Z · score: 1 (1 votes) · EA · GW Also just came across this claim in this paper: In the nationally representative UK Household Longitudinal Survey ("Understanding Society"), for example, average life satisfaction, measured on a scale from one to seven whereby higher values denote higher wellbeing, was not significantly higher in 2016 than in 1996 (5.3 vs. 5.2), despite large rises in real incomes. Couldn't quickly chase down source data up through 2016--best I could find was this through 2008. Comment by cole_haus on Growth and the case against randomista development · 2020-01-28T01:04:51.779Z · score: 9 (4 votes) · EA · GW How Poverty Ends: The Many Paths to Progress—and Why They Might Not Continue is from Abhijit V. Banerjee and Esther Duflo (the recent econ Nobel Laureates who won for their RCT work) and I think can reasonably be read as a response to the criticisms of Lant Pritchett (probably the most vocal advocate of the line of thinking this post represents). Key excerpts: Economists, ourselves included, have spent entire careers studying development and poverty, and the uncomfortable truth is that the field still doesn’t have a good sense of why some economies expand and others don’t. There is no clear formula for growth. What, then, are policymakers left with? There are some things clearly worth avoiding: hyperinflation; extremely overvalued fixed exchange rates; communism in its Soviet, Maoist, or North Korean varieties; the kind of total government chokehold on private enterprise that India had in the 1970s, with state ownership of everything from shipyards to shoe factories. But this is not particularly helpful advice today, given that hardly anyone is reaching for such extreme options anymore. There simply is no accepted recipe for how to make poor countries achieve permanently high growth. Even the experts seem to have accepted this. In 2006, the World Bank asked the economist Michael Spence to lead a commission on economic growth. In its final report, the group recognized that there are no general principles for growth and that no two instances of economic expansion are quite alike. Easterly described their efforts in less charitable terms: “After two years of work by the commission of 21 world leaders and experts, an 11-member working group, 300 academic experts, 12 workshops, 13 consultations, and a budget of4m, the experts’ answer to the question of how to attain high growth was roughly: we do not know, but trust experts to figure it out.”

If there is a common thread, it is that the fastest growth appears to come from reallocating poorly allocated resources—that is, putting capital and labor toward their most productive use. But eventually, the returns from that process diminish, at which point countries need to find a new strategy for combating poverty.

One very real danger is that in trying to hold on to fast growth, countries facing sharply slowing growth will veer toward policies that hurt the poor now in the name of future growth. In a bid to preserve growth, many countries have interpreted the prescription to be business friendly as a license to enact all kinds of anti-poor, pro-rich policies, such as tax cuts for the rich and bailouts for corporations.

Looking back, it is clear that many of the important successes of the last few decades were the result not of economic growth but of a direct focus on improving particular outcomes, even in countries that were and have remained very poor. The under-five mortality rate, for example, has fallen drastically across the world, even in some very poor countries whose economies have not grown particularly fast. Credit goes mostly to policymakers’ focus on newborn care, vaccination, and malaria prevention.

This is patient work: spending money by itself does not necessarily deliver real education or good health. But unlike with growth, experts actually know how to make progress. One big advantage of focusing on clearly defined interventions is that these policies have measurable objectives and therefore can be directly evaluated. Researchers can experiment with them, abandon the ones that don’t work, and improve the ones that do.

Comment by cole_haus on Optimal population density: trading off the quality and quantity of welfare · 2020-01-24T01:32:57.920Z · score: 1 (1 votes) · EA · GW

Maybe more standing variation and population resilience which could have indirect effects on welfare over the long-term?

Comment by cole_haus on Optimal population density: trading off the quality and quantity of welfare · 2020-01-24T01:26:53.165Z · score: 1 (1 votes) · EA · GW

I'm struggling to think of non-human-animal-relevant examples of this, but it seems in principle possible to have welfare effects which depend on total population rather than density. Ideas and technological innovation are (roughly) an example of this in humans. That is, more total humans means more freely shareable ideas/innovations means more welfare. (See e.g. Population growth and technological change: One million BC to 1990.)

Thought I'd just throw it out there in case someone can think of a way to make this relevant to animals.

Comment by cole_haus on Potential downsides of using explicit probabilities · 2020-01-24T01:07:35.768Z · score: 2 (2 votes) · EA · GW

Some related things that come to mind:

• Challenges to Bayesian Confirmation Theory outlines some conceptual potential issues arising from the use of explicit probabilities in a Bayesian framework.
• Gerd Gigerenzer likes to claim that "fast and frugal" heuristics often just perform better than more formal, quantitative models. These claims can be linked to the bias-variance tradeoff and extreme priors.
• The optimizer's curse can be generalized to the satisficer's curse. This generalization doesn't obviously seem to differentially affect explicit probabilities though.
Comment by cole_haus on Growth and the case against randomista development · 2020-01-18T01:11:15.572Z · score: 6 (6 votes) · EA · GW

Having looked at the paper now, I definitely have a different take as to how definitive it is. My maximally contrarian take would be that it's a non-systematic review in which many (most?) of the works reviewed are in favor of an important causal link running from health to income. I do agree that the overall macro-scale evidence is weak (which is distinct from strong evidence of a weak effect), but this is exactly why people like RCTs over national development! Causal inference at a macro scale is hard!

(Health and Economic Growth: Reconciling the Micro and Macro Evidence also looks like a good source that I'll look at.)

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T22:31:53.896Z · score: 1 (1 votes) · EA · GW

It's not binary, though.

Yup, agreed.

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T22:26:40.840Z · score: 1 (1 votes) · EA · GW

The link I was thinking of is that migration loans are relevant for urbanization, agglomeration effects, and generally the distribution and density of humans (which seem like they fit into human geography--"Human geography attends to human patterns of social interaction, as well as spatial level interdependencies"--and urban geography).

Deworming doesn't change the geography of a place in itself but it mediates the impact of that geography on humans/the economy/society (supposing that we consider the disease environment a part of geography similar to the way we might consider forestation and the endowment of domesticable animals as geographic factors).

Similarly, fertilization mediates the impact of poor soils which are prevalent in equatorial regions.

Not sure if that makes the examples and their connection to geography more scrutable.

What are you thinking of when you emphasize geography as a determinant of growth? Stuff like shipping access and which countries are neighbors?

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T22:15:17.713Z · score: 13 (7 votes) · EA · GW

The Easterlin paradox notwithstanding, economic growth does buy you a lot of subjective wellbeing improvement in a country.

I might not be understanding you, but it seems like this tries to smuggle in causation and assume away the problem. As I see things, there are two conflicting pieces of correlational evidence:

• Cross-country regressions show strong correlation at a point in time between income and SWB (what the post highlights)
• Time series regressions within countries show a weak correlation between income and SWB (Easterlin paradox)

I don't currently know of a fully convincing resolution of this conflict, but the second correlation actually seems a bit more central for the question of the causal effect of growth over time on SWB.

It would be interesting to explore how far increasing growth in a country would improve subjective wellbeing in LMICs.

Some analysts assert that in less developed countries happiness and economic growth are positively related “up to some point,” beyond which the association tends to become nil, but time series data do not support this view. The most striking contradiction is China where, despite a fourfold multiplication in two decades in real GDP per capita from a low initial level, life satisfaction has not improved.

(For the record, I would be surprised if the Easterlin paradox turned out to 100% correct and rising national income over time had 0 positive effect on subjective well-being. But I am significantly uncertain about this, could imagine the effect being quite substantial, and knowing the magnitude of the effect seems very important.)

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T21:52:35.489Z · score: 1 (1 votes) · EA · GW

I updated in the direction that health and education are not that important for growth at least for very poor countries

Yeah, I will have to look into this perspective more.

But it's a different question on whether we can do anything about that by ramping up health spending and ameliorate these differences and whether that's important for growth.

I do think it's an open question though.

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T21:47:53.324Z · score: 3 (2 votes) · EA · GW

Yup, agreed that none of the linked things are on growth per se. I just think the link to the systemic change objection is useful because it gives hints as to what problems there might be with the growth-focus argument, how people are likely to react to the growth-focus argument, which arguments are persuasive, etc.

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T21:36:36.921Z · score: 1 (1 votes) · EA · GW

Yup, agree that the argument I outline is not definitive and thoughtful work in this area is worthwhile. I think I may be more pessimistic on the politics aspect (i.e. I may think it's a more tightly-binding constraint and harder for outsiders to work on), but my sense of that is kind of inchoate and not worth much at the moment.

Comment by cole_haus on Making decisions when both morally and empirically uncertain · 2020-01-17T04:34:06.456Z · score: 2 (2 votes) · EA · GW

not really understanding how ordinal moral theories are really meant to work

Yeah, I think this is where I'm at too. It seems inescapable that ordinal preferences have cardinal implications when combined with empirical uncertainty (e.g. if I prefer a 20% chance of A to an 80% chance of B, that implies I like A at least four times as much). The only choice we really have is whether the corresponding cardinal implications are well-formed (e.g. Dutch bookable). The best distinctions I can come up with are:

• In a purely deterministic world without lotteries, there wouldn't be an obvious mechanism forcing the cardinalization of ordinal preferences. So their overlap is only a contingent feature of the world we find ourselves in. (Though see A Theory of Experienced Utility and Utilitarianism for an alternate basis for cardinalization.)
• Ordinal preferences only specify a unique cardinalization in the limit of an infinite sequence of choices. Since we aren't likely to face an infinite sequence of choices any time soon, they're distinct in practice.

P.S. Thanks for the Tarsney link. I have it open in a tab and should get around to reading it at some point.

Plus there's the roadblock of me not having in-depth understanding of how the vNM utility theorem is meant to work.

Not sure if it'll help but I have a short explanation and interactive widget trying to explain it here.

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T02:07:50.346Z · score: 9 (7 votes) · EA · GW

China’s growth acceleration from 1977 onwards produced $14 trillion NPV in cumulative economic output. Thus, if the only thing the economics profession achieved in 50 years was to increase by 4 percentage points the probability that the Chinese government shifted to this new economic strategy, then it would have had greater economic benefits than the Graduation approach. I think this argument equivocates between the probability of any reform and the probability of a particular reform. Because the reform policy was academic-economist-inflected, denying the influence of economists sounds silly. But I think we should instead think about two separate chunks: • the wedge between the status quo growth trajectory and the reform trajectory that would have obtained had Chinese reformers only had e.g. 1960s economics knowledge • the wedge between the 1960s-style-reform trajectory and the actual reform trajectory Economists since the 1960s get 100% of the credit for the second wedge, but I think it's plausibly extremely small (especially since Chinese reforms didn't hew particularly closely to economic orthodoxy in the details). Economists since the 1960s only get any credit for the first wedge insofar as they were the impetus for major economic reforms. My limited knowledge of Chinese economic history suggests that this probability could easily be very small. Phrased differently, we can model this as: Without post-1960s economists: • Status quo had X% chance of continuing • Reform had (1-X)% chance of happening and represented a$Z gain

With post-1960s economists:

• Status quo had X-A% chance of continuing
• Reform had (1-X)+A% chance of continuing and represented a \$Z+B gain

I can easily imagine A and B both being quite small but the original presentation disguises that.

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T01:12:05.020Z · score: 55 (28 votes) · EA · GW

Here's a central argument against focusing on growth per se that I find fairly plausible:

Obviously terrible growth-related policies are at historic lows. Our ability to produce more detailed/refined policy prescriptions is weak (see Pritchett's acknowledgement of the lost decades and the transition depression). In fact, many of the greatest successes of development (China, Singapore, etc.) defied the economic orthodoxy in the details. Rather, they implemented policies that were tailored to and required deep understanding of local conditions. The key barrier to increased economic growth is not the absence of knowledge or advocacy but mundane implementation issues and the indifference or antipathy of the relevant political actors.

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T00:54:43.299Z · score: 48 (27 votes) · EA · GW

Wild speculation:

I think one reason this area may get less attention in EA is that if you're willing to sign up for high-risk high-return scenarios that are more theory-driven and less retrospective-data-driven (like economic growth), you're also more sympathetic to long-termist areas like x-risk. And once you're comparing x-risk to economic growth, there's no guarantee that growth wins.

In other words, I think economic growth may be competing against x-risk--not RCTs--among EAs.

(Though certain ethical views may argue against long-termist interventions like x-risk reduction. A focus on economic growth may be the best fit for people that are "epistemically permissive" but "ethically conservative", if that makes sense.)

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T00:47:49.838Z · score: 5 (4 votes) · EA · GW

RD has moved in an entirely different direction. Instead of replicating this success, it asks: among interventions that we can test with RCTs, what is most impactful? In the wake of the period with by far the greatest progress in human welfare of all time, this change in strategy is difficult to justify.

I think one possible explanation (I've not heard this anywhere explicitly; it's just me making things up.) that I find moderately persuasive is:

Development RCTs rose to prominence in the wake of the '“lost decades” in Latin America and the “transition depression” in some (not all) former Soviet dominated countries'. These failures and the broader (perceived) failures of the Washington Consensus provoked a crisis of confidence in economic policy prescriptions. In particular, large-scale economic theory is hard to make accountable to empirical evidence. RCTs shore up this epistemic weakness--they allow economists to test their theories against reality and build up more certain knowledge which can perhaps eventually be applied to big questions of national development. Without RCTs, economic prescriptions are much more reliant on theory (even non-RCT causal inference from empirical data is more theory-laden) and it doesn't seem great to cut out (almost) one whole category of evidence. (See What randomization can and cannot do for a good discussion of the interplay of theory and RCTs.)

I think epistemics is plausibly a crux for in the randomista vs big national development debate.

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T00:34:24.230Z · score: 6 (6 votes) · EA · GW

This should lead us to be sceptical about RD. Growth is arguably the major driver of human progress, but proponents of RD rarely argue that the interventions that they recommend do increase growth.

To the extent that this is true, I think there are pretty benign possible explanations:

• The data on growth (e.g. GDP) and RD interventions typically operate at different scales. Even if GiveDirectly substantially increases the long-term growth in a village, that's not going to show up in national aggregate data.
• Making empirical growth claims requires data accumulated over time. The randomista trend is pretty recent. Using Randomized Controlled Trials to Estimate Long-Run Impacts in Development Economics mentions that looking at long-run impacts will become more and more possible as more early RCTs reach the age of maturity.
Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T00:24:15.577Z · score: 1 (4 votes) · EA · GW

Thus, many RCT-backed interventions do not seem to explain much of the cross-national variation in GDP per capita. What does? There are a range of factors including:

• Growth-friendly policies
• Geography
• Natural resources
• Human capital
• Culture

I'm confused here. It seems like there are examples of RCTs addressing at least:

• Geography: It seems like No Lean Season, Duflo's fertilizer nudges, and deworming (edit: I try to make the link between these and geography more obvious in a grandchild comment.) are all examples of RCTs targeting various aspects of geography depending on how you want to interpret the term.
• Human capital: Education and health are clear focus areas for development RCTs.
• Culture: It seems likely that many RCTs have flow-through effects on culture.

Is the argument not that RCTs can't address things in these categories but that they aren't good at it?

Comment by cole_haus on Growth and the case against randomista development · 2020-01-17T00:10:16.454Z · score: 5 (5 votes) · EA · GW

The reason these things are unlikely to be the best way to increase growth is that they play no role in the causal story of the huge differences in GDP per capita across space and time.

...

But given the story above, it would be very surprising if this was the case: differences in rates of deworming explain a miniscule fraction of the variation in individual economic outcomes across the world. No-one argues that deworming is among the top 1000 causes of the huge economic transformation documented above.

These seem like clearly insufficient arguments. They strike me as analogous to:

Don't focus on fixing your broken leg. What separates the top 20 bicyclists from the next 40 is the kind of bike they ride. Focus on getting the right bike.

Different circumstances apply in different places and focusing on and importing salient characteristics of the winners is not likely to be successful because it's almost never the case in complex domains like this that a single factor is sufficient for success on its own.

More concretely, while no one argues that deworming is a top cause of the huge economic transformation, plenty of people argue that disease environment/burden (or geography more broadly) is an important cause of differential economic success. For example, Historical Development references arguments that the tsetse fly had substantial impacts on Africa's path of development ("by inhibiting the development of intensive agriculture using draft animals, resulted in lower populations, less urbanization, and less state development").

Comment by cole_haus on Growth and the case against randomista development · 2020-01-16T23:46:30.232Z · score: 5 (4 votes) · EA · GW

It seems like one of the key implicit claims in the post is that growth effects are better/more important than level effects (e.g. The post says "Moreover, the vast majority of proponents of RD do not tackle the question of whether the interventions they assess increase economic growth." which is true, but RD proponents often focus on level effects) . I think it would be good to state and argue for this point explicitly.

Relatedly, I think the anti-RD perspective advocated in this post require the claim that level effects don't affect growth rates. If boosting someone's assets or income leads to a persistent increase in their income growth, the RD-caused level effect also gets the growth benefits this post argues for. The low-level equilibrium trap is a pretty popular model which describes just this dynamic.

Comment by cole_haus on Growth and the case against randomista development · 2020-01-16T23:30:28.181Z · score: 1 (1 votes) · EA · GW

As discussed above, there are also good reasons to believe that increased GDP per capita causes many of these increases in welfare.

Can you help clarify what the causal evidence is? I don't really see any non-correlational evidence in the preceding section. (I'm assuming that's what this sentence is emphasizing given that the opening sentence in the preceding paragraph is "The foregoing arguments show that GDP per capita is strongly correlated with many objective and subjective measures of welfare.") I think causality and evidence of it is actually pretty central to the debate---one of the chief advantages of RCTs is the ability to make causal claims. You can of course use things like instrumental variables and natural experiments to make causal claims about national development-type interventions but that's pretty hard (see e.g. The Skeptics Guide to Institutions - Part 1).

The claim that increased GDP per capita causes other beneficial changes seems plausible to me, but I also don't have too much trouble half-convincing myself that all these things are driven by common causes (e.g. it's pretty easy for me to believe (based on a substantial literature) that increased social trust leads to both increased GDP per capita and increased welfare).

Comment by cole_haus on Growth and the case against randomista development · 2020-01-16T23:17:40.750Z · score: 20 (11 votes) · EA · GW

GDP per capita is also correlated with self-reported life satisfaction

It seems useful to point out (because I presume not all readers will know this) that subjective well-being is often divided into positive affect, negative affect, and life satisfaction. Of the three, life satisfaction tends to be the most tightly correlated with measures like GDP per capita. So the correlation between GDP per capita and life satisfaction isn't quite as definitive a statement about subjective well-being as it might naively appear.

Comment by cole_haus on Growth and the case against randomista development · 2020-01-16T23:12:44.722Z · score: 5 (3 votes) · EA · GW

There's tons of back-and-forth on the Easterlin paradox ("The paradox states that at a point in time happiness varies directly with income both among and within nations, but over time happiness does not trend upward as income continues to grow."), but if we're talking about national development over time (i.e. policy-oriented growth) this seems of pretty central relevance. In particular, it's plausibly more of a problem for growth-oriented national development than for RD because one plausible explanation for the paradox is the importance of positional goods (I care not just about my absolute income but how it compares to my neighbors or how it compares to my expectations which are influenced by neighbors). If the whole country (everyone in my comparison class) grows their GDP per capita at exactly the same rate, everyone's position is unchanged; if the poorest are targeted by RD, their positions can be improved with minimal harm to others.

(This shades into inequality as an important consideration beyond GDP as the post mentions, but it's not quite the same.)

Comment by cole_haus on Growth and the case against randomista development · 2020-01-16T22:59:56.240Z · score: 12 (8 votes) · EA · GW

In spite of this, to our knowledge, this question has received no publicly published attention from the EA community.

There are a few claims like this in the post. I think there is prior related work. Narrowly, a recent example is Effective Altruism and International Trade. More broadly, I think there are strong links between the line of debate in this post and the perennial "systemic change objection" (as alluded to by jonathanpaulson in another comment). Recent stuff on the systemic change objection includes e.g. Effective Altruism and Systemic Change and Some personal thoughts on EA and systemic change.

(I don't want to get into a full explanation/discussion of the analogies between the systemic change discussion and the growth objection in this comment but just for the sake of clarity:

• Both about relatively small and certain impacts vs large and more speculative impacts
• Both about interventions with substantial empirical evidence vs more theory-driven interventions

)

Hopefully pointing out these related discussions comes across as a helpful pointer to further thinking and not a "Gotcha!".

Comment by cole_haus on Growth and the case against randomista development · 2020-01-16T22:47:23.257Z · score: 10 (5 votes) · EA · GW

Improving health is not the best way to increase growth.

I'll look through the linked paper, but I'd be surprised if one paper is enough to convince me of the spirit of this claim (which I take to be not just that health is not the best but not even especially good or worth targeting). The impression I get is that consensus in development economics is that human capital interventions (e.g. education and health) are very well-regarded. For example Using Randomized Controlled Trials to Estimate Long-Run Impacts in Development Economics says:

"In this section, we assess the evidence from the emerging body of literature that exploits RCTs to estimate long-run impacts of development interventions. One pattern that emerges from the handful of existing studies is that human capital interventions appear to be particularly effective at boosting long-run economic outcomes."

Comment by cole_haus on Growth and the case against randomista development · 2020-01-16T22:39:29.979Z · score: 3 (2 votes) · EA · GW

Thanks for writing this up! I'm very interested in this area (was/am actually working on something related) and open/sympathetic to the overall claim. I'll make my substantive comments separately (to work better with comment threading/nesting).

Comment by cole_haus on Making decisions when both morally and empirically uncertain · 2020-01-13T23:57:33.368Z · score: 2 (2 votes) · EA · GW

Maybe worth noting that ordinal preferences and a probability distribution over empirical outcomes (the setup in "BR under empirical uncertainty") are used to generate cardinal preferences in the vNM utility theorem.

Comment by cole_haus on 2019 AI Alignment Literature Review and Charity Comparison · 2020-01-13T21:24:39.307Z · score: 15 (6 votes) · EA · GW

I'd guess it's because it's very hard to apply a GiveWell approach (i.e. explicit quantitative modeling based on a substantial body of empirical evidence) to many long-termist orgs (whose impacts will often definitionally be unknown for the foreseeable future, and around which there is large and pervasive uncertainty). Open Philanthropy describes an evaluation methodology that seems more suited to long-termist orgs.

Comment by cole_haus on Space governance is important, tractable and neglected · 2020-01-13T00:10:53.876Z · score: 1 (1 votes) · EA · GW

This isn't central to the topic, but Paul Krugman's The Theory of Interstellar Trade is fun and a bit related.

Comment by cole_haus on Disadvantages of the measures · 2020-01-12T22:35:33.225Z · score: 9 (7 votes) · EA · GW

Metrics can distort info by:

• Measuring the most easily measurable
• Measuring the simple when the desired outcome is complex
• Measuring inputs rather than outcomes
• Degrading information quality through standardization

Metrics can be gamed by:

• Creaming (e.g. surgeons only taking on the easiest surgeries to improve their stats)
• Improving numbers by lowering standards
• Improving numbers through omission or distortion of data
• Cheating

Negative consequences of metrics include:

• Goal displacement through diversion of effort to what gets measured.
• Promoting short-termism.
• Costs in employee time.
• Diminishing utility.
• Rule cascades. In an attempt to staunch the flow of faulty metrics through gaming, cheating, and goal diversion, organizations institute a cascade of rules.
• Rewarding luck. Measuring outcomes when the people involved have little control over the results is tantamount to rewarding luck.
• Discouraging risk-taking.
• Discouraging innovation.
• Discouraging cooperation and common purpose.