Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

The Effectiveness of Developing World Aid

19 Post author: multifoliaterose 12 September 2010 09:56PM

Several Less Wrong posters [1], [2], [3] have cited the interview with James Shikwati titled "For God's Sake, Please Stop the Aid!" as evidence that Western aid to Africa is actively destructive to Africa. According to the wikipedia page on James Shikwati:

Jeffrey D. Sachs, a Columbia University professor who is a leading aid advocate, calls Mr. Shikwati’s criticisms of foreign assistance “shockingly misguided” and “amazingly wrong.” “This happens to be a matter of life and death for millions of people, so getting it wrong has huge consequences,” Mr. Sachs said.

I think it's important for those interested in the question of whether developing world aid is effective to look to those who can point to formal studies about the effectiveness of African aid rather than basing their judgments on quotes from individuals whose opinions may very well have been heavily skewed by selection bias and/or driven by ideological considerations which have nothing to do with the available evidence.

Engaging with the evidence in detail is a very time-consuming task and one beyond the scope of this blog entry. I will however quote various experts with links to useful references.

Divided Views On Overall Impact of Developing World Aid:

I've found Paul Collier to be apparently even-handed. Readers interested in studying the the effectiveness of developing world aid may like to study Paul Collier's papers on the subject. Paul Collier summarizes his views in his recent book titled The Bottom Billion.

The left seems to want to regard aid as some sort of reparations for colonialism. In other words, it's a statement about the guilt of Western society, not about development. In this view, the only role for the bottom billion is as victims: they all suffer from our sins. The right seems to want to equate aid with welfare scrounging. In other words, it is rewarding the feckless and so accentuating the problem. Between these two there is a thin sliver of sanity called aid for development. It runs something like this. We used to be that poor once. It took us two hundred years to get where we are. Let's try to speed things up for these countries.

Aid does tend to speed up the growth process. A reasonable estimate is that over the last thirty years it has added around one percentage point to the annual growth rate of the bottom billion. This does not sound like a whole lot, but then the growth rate of the bottom billion over this period has been much less than 1 percent per year - in fact it has been zero. So adding 1 percent has made the difference between stagnation and severe cumulative decline. Without aid, cumulatively the countries of the bottom billion would have become much poorer than they are today. Aid has been a holding operation to prevent things from falling apart.

[...]

...unlikely as it seems, what aid agencies have been doing has added a whole lot of value to the financial transfer. Given the bad public image of aid agencies and horror stories such as the hospital project I described above, this is hard to believe, but there it is.

[...]

Aid, however, is not the only answer to the problems of the bottom billion. In recent years it has probably been overemphasized, partly because it is the easiest thing for the Western world to do and partly because it fits so comfortably into a moral universe organized around the principles of sin and expiation. That overemphasis, which comes from the left, has produced a predicable backlash from the right. Aid does have serious problems, and more especially serious limitations. Alone it will not be sufficient to turn the societies of the bottom billion around. But it is part of the solution rather than part of the problem. The challenge is to complement it with other actions.

An economist who is skeptical of Collier's analysis William Easterly, the author of The White Man's Burden: Why the West's Efforts to Aid the Rest Have Done So Much Ill and So Little Good. For those who are interested in the topic of the effectiveness of developing world aid, Easterly's papers are also worth taking a look at. In an article for Boston Review, Easterly writes:

Collier’s work is built on deeply problematic statistical analysis. Valid statistical results must meet stringent conditions. The usual standard for labeling a result “significant” is that it could have occurred by chance only one out of twenty times, assuming a statistical exercise is run only once. An unfortunately all-too-common practice called “data mining” involves running twenty statistical exercises and then reporting only the one that produces a “significant” result (which will have happened by chance). Collier comes close to admitting that he does exactly that.

[...]

Remarkably enough, Collier puts the burden of proof on non-intervention. “At some point,” he writes, “doubt becomes an excuse for inaction, while the problems of insecurity remain real enough.” Elsewhere Collier alludes to his doubters as “professional skeptics.” Actually, doubt is a superb reason for inaction. If being a professional skeptic entails scrutinizing the logic, the assumptions, and the evidence base and finding them all invalid when they do not meet normal academic standards, I plead guilty.

My overall impression is that there's a fair amount of controversy as to whether African aid has increased economic growth in Africa. Different economists have different views and the evidence available does not seem sufficiently robust to support a confident belief that the net effect of aid has been positive or negative.

Much less controversial is the view that the best health Western interventions can and do systematically improve health in Africa.

The Case of Health:

In his article titled Can The West Save Africa? Easterly writes

Health is an even more clear success story than education in Africa, as child mortality has improved dramatically over time ... There are well known and striking donor success stories, like the elimination of smallpox, the near-eradication of river blindness and Guinea worm, the spread of oral rehydration therapy for treating infant diarrheal diseases, DDT campaigns against malarial mosquitoes (although later halted for environmental reasons), and the success of WHO vaccination programs against measles and other childhood diseases. The aid campaign against diseases in Africa (known as vertical health programs, see discussion below) is likely the single biggest success story in the history of aid to Africa.

In this case, the clear verdict of the case studies is probably a lot more helpful than the aggregate stylized facts, aggregate econometrics, or REs. Under-five mortality fell dramatically in Africa, but it fell by somewhat less than in other developing countries. We ideally need to parcel out factors such as Africa’s lower growth (although the effect of growth on health is controversial), different disease ecology (for example, malaria is much more of a problem in Africa than any other region), other factors, and aid, not to mention finding an identification strategy to assess causal effects of aid; no such aggregate econometric efforts have been notably successful. Even with econometric support unavailable, perhaps Africa’s health performance is impressive after all given its lower growth and its more difficult disease ecology, which is consistent with the important role for aid shown by the case studies.

According to the GiveWell page titled Why do we look for charities implementing proven programs:

The most successful projects have been in the area of health and include such large-scale successes as the eradication of smallpox and the dramatic reduction of infant mortality in Africa (see our developing-world health overview). For a full list of health programs that have been rigorously shown to save lives and reduce suffering, see our summary of proven health programs.

There are a number of examples of ways in which well-intentioned projects have failed to achieve desired results. Building wells has often failed to reduce water-related illness (detailed analysis here); agriculture programs in Africa have failed to increase crop yields; programs providing textbooks and other supplies have not raised students' test scores, and many other developing-world education programs have weak, if any, evidence of success.

Charities working on improving health in the developing world have variable effectiveness. It's plausible that by donating to one of GiveWell's top-rated charities one can have a substantially stronger positive effect than the one associated to a random such charity. In Charity Isn't About Helping? Holden says:

One person who’s more critical of charity than we are or than David Hunter is is the economist Robin Hanson. He has stated that “charity isn’t about helping”.

[...]

What response can the nonprofit sector marshal to arguments like this? I must say that, in fact, much of the nonprofit sector fits incredibly better into Prof. Hanson’s view of charity as “wasteful signaling” than into the traditional view of charity as helping.

[...]

Perhaps ironically, if you want a good response to Prof. Hanson’s view, I can’t think of a better place to turn than GiveWell’s top-rated charities. We have done the legwork to identify charities that can convincingly demonstrate positive impact. No matter what one thinks of the sector as a whole, they can’t argue that there are no good charitable options - charities that really will use your money to help people - except by engaging with the specifics of these charities’ strong evidence.

Valid observations that the sector is broken - or not designed around helping people - are no longer an excuse not to give.

Because our Bayesian prior is so skeptical, we end up with charities that you can be confident in, almost no matter where you’re coming from.

A Note on Malthusian Problems:

Those who are concerned about possible future Malthusian problems attached to saving lives should see Holden's email to the GiveWell research mailing list titled Population growth & health, the video linked therein, and papers by the speaker Hans Rosling. I presently believe that while it's possible that saving lives in the developing world does more harm than good on account of Malthusian problem, this is fairly unlikely and the expected value of saving lives in the developing world is strongly positive. Of course, my belief is subject to change with incoming evidence.

The Giving What We Can (GWWC) Myths About Aid page provides a suggestion for donors who are concerned about future Malthusian problems:

...there are many ways that you can greatly improve the lives of thousands of people who live in extreme poverty without significantly extending these lives. For example, you could cure people of blindness, or of neglected tropical diseases, which cause significant hardship but have only a small effect on mortality. Alternatively, you could donate to groups who promote family planning in developing countries, directly fighting population growth. Those who think that overpopulation is so bad that we should let people suffer and die rather than risk saving their lives, must also think it is important enough that they should donate money to groups that directly fight it.

GWWC and GiveWell differ in that GWWC's top recommended charities are Schistosomiasis Control Initiative and Deworm the World which focus on neglected tropical diesease whereas GiveWell does not recommend these organizations. Holden explains GiveWell's position in Neglected Tropical Disease charities: Schistosomiasis Control Initiative, Deworm The World.

A Note on Overcorrecting Bias:

Many people's initial naive reaction to developing world aid is that it's a very good idea. This was certainly my own reaction as a child when I learned of Unicef. As Eliezer suggests in Can't Say No To Spending, there's a natural bias in favor of saying "yes" when asked to donate money help poor people - saying 'no' feels cold-hearted. Reading an author like Shikwati can dispel this bias by making possible unintended negative consequences salient, but often at the cost of giving rise to a new bias against developing world aid. Reading Sachs' remarks on Shikwati quoted in the introduction of this article can dispel this bias at the cost of introducing a new bias in favor of developing world aid. But Sachs' own position has garnered seemingly valid criticism from William Easterly and others - learning of this introduces a bias against Sachs and his views and in favor of Easterly and his views. There's the usual issue of there being a halo effect as described in Yvain's excellent article titled The Trouble with "Good" - when person X debunks person Y's apparently erroneous claim, this makes person X look unwarrentedly superior to person Y overall.

It's difficult to know who to trust when ostensible experts disagree, any of whom may be exhibiting motivated cognition or even engaging in outright conscious self-serving deception. Nevertheless, one can reasonably can hope to arrive at a fairly good epistemological state by:

  1. Reading representatives of a wide variety of perspectives
  2. Paying special attention to those experts who are willing to engage with differing perspectives in detail
  3. Being careful to keep in mind that somebody may have valid points which are worthy of consideration independently of whether their general thesis is correct
  4. Paying special attention to points of common agreement among experts

Habits (1)-(4) are conducive to converging on a relatively accurate epistemological position on a given matter.


Acknowledgment: Thanks to Carl Shulman for useful references and discussion about the subject of this article.

Comments (51)

Comment author: Vladimir_M 13 September 2010 08:12:15PM *  12 points [-]

multifoliaterose:

I think it's important for those interested in the question of whether developing world aid is effective to look to those who can point to formal studies about the effectiveness of African aid rather than basing their judgments on quotes from individuals whose opinions may very well have been heavily skewed by selection bias and/or driven by ideological considerations which have nothing to do with the available evidence.

Trouble is, most of these studies are far from immaculate when it comes to the ideological or even career interests of their authors. In areas of research whose conclusions cannot be separated from their political and ideological implications, there is no rational reason to believe that the mainstream academic opinion is on the right track at all.

For example, you quote Jeffrey Sachs as someone whose expert authority is supposed to be a strong blow against Shikwati. But what do you expect from someone who has the career track and affiliations like Sachs when faced with such arguments? To say openly that his life work for which he's reaped status, fame, and power might have been, to a significant degree, a pernicious scam?

It reminds me of the dialogue from "Blackadder" when Edmund starts getting skeptical about his (16th century) doctor who insists that leeches are a cure-all:

Doctor: You know the leech comes to us on the highest authority?
Edmund: Yes. I know that. Dr. Hoffmann of Stuttgart, isn't it?
Doctor: That's right, the great Hoffmann!
Edmund: Owner of the largest leech farm in Europe.
Doctor: Yes!

Comment author: multifoliaterose 13 September 2010 09:14:23PM *  1 point [-]

Trouble is, most of these studies are far from immaculate when it comes to the ideological or even career interests of their authors. In areas of research whose conclusions cannot be separated from their political and ideological implications, there is no rational reason to believe that the mainstream academic opinion is on the right track at all.

This seems to me to be true for some topics.

I think that the best thing to do in such circumstances is to adopt a position of agnosticism about the issue at hand (as I do with respect to whether the overall impact of aid has been positive or negative).

As for the special case of health interventions, I haven't seen any expert authority mount an apparently credible case against the idea that health aid has had a positive impact. Have you?

For example, you quote Jeffrey Sachs as someone whose expert authority is supposed to be a strong blow against Shikwati.

My quotation of Sachs was intended to highlight the fact that experts have wildly differing views on the subject of whether aid has been good or bad and that it's important not to take quotes from any one individual (including Sachs) too seriously.

But what do you expect from someone who has the career track and affiliations like Sachs when faced with such arguments? To say openly that his life work for which he's reaped status, fame, and power might have been, to a significant degree, a pernicious scam?

I agree that the issue that you allude to is a serious concern. See my discussion in the final section of my top level post titled "A Note on Overcorrecting Bias."

Comment author: utilitymonster 17 September 2010 07:02:36PM 2 points [-]

On this point, it is noteworthy that international health aid eliminated small pox. According to Toby Ord, it is estimated that this has prevented over 100 million deaths, which is more than the total number of people that died in all wars in the 20th century. If you assumed that all of the rest of international health aid achieved nothing at all, this single effort would make the average number of dollars per DALY achieved by international health aid better than what the British Government achieves.

Comment author: taw 22 September 2010 09:41:43PM 0 points [-]

You can always pick a reference class which supports any conclusion you want.

You could just as plausibly claim that international aid mostly propped up various third world dictators and fueled local wars (no matter what "aid money" was for, government could always shift money it would otherwise need to spend on that area into buying weapons or beating up dissidents instead), leading to economic and civilizational stagnation, and over 100 million deaths which would otherwise not have happened.

Or you could categorize reality into reference classes any other way.

Comment author: PhilGoetz 13 September 2010 07:53:40PM *  7 points [-]

I think it's important for those interested in the question of whether developing world aid is effective to look to those who can point to formal studies about the effectiveness of African aid rather than basing their judgments on quotes from individuals whose opinions may very well have been heavily skewed by selection bias and/or driven by ideological considerations which have nothing to do with the available evidence.

Do you think that people who believe aid harms are more likely to be driven by ideological considerations? I'd expect the opposite, because ideologies that say aid is good are very popular.

Engaging with the evidence in detail is a very time-consuming task and one beyond the scope of this blog entry. I will however quote various experts with links to useful references.

But... you just said... Wait - you mean you believe there are humans whose opinions are not heavily skewed by selection bias and/or driven by ideological considerations which have nothing to do with the available evidence?

I presently believe that while it's possible that saving lives in the developing world does more harm than good on account of Malthusian problem, this is fairly unlikely

The demographics of Africa are currently crazy skewed towards kids. Look at what the CIA Factbook says about median age. For almost every country in Africa except Libya, Angola, and Egypt, the median age is 15-20. For other countries, it's 30-40. The countries outside Africa with Africa-like median ages are: Afghanistan, Bolivia, the Gaza strip, Guatemala, Pakistan, the West Bank, and Yemen.

The evidence is overwhelming <EDIT>not really, see below</EDIT> that having lots of kids is strongly correlated with poverty, civil unrest, and war. Causation probably runs in both direction. But imagine trying to run a country when most of your citizens are in their teens. Is it a coincidence that the countries that are a nexus for terrorism, like Somalia, Afghanistan, the Gaza strip, the West Bank, Yemen, and Pakistan, have very low median ages? I don't think so. These countries have too many children for the adults to control. And lots of them have AK-47s.

In short: A sudden decrease in child mortality <EDIT>usually</EDIT> causes terrorism and civil war.

This brings to mind the chapter in Freakonomics arguing that the introduction of abortion to the US caused the rapid decline in violent crime about 20 years later.

Look at this graph on the demographics of suicide bombers. 82% were under age 25. 98.7% were under age 35. (Suicide bombers are much more likely to be single than married; so there is probably an age x chance of being married variable to factor out - assuming "being single" has causal power.)

(I also note that 77% of those with known educational histories went to college - while about 20% of all men in the West Bank go to college, based on population and enrollment figures and the assumption that zero women attend college in the West Bank.)

Comment author: PhilGoetz 16 September 2010 09:19:57PM *  5 points [-]

The evidence is overwhelming that having lots of kids is strongly correlated with poverty, civil unrest, and war.

I repent of this statement!

The sample size is large, but the countries in Africa, and those in the Middle East, are too highly-correlated within their respective groups in many other ways, including climate, culture, and genetics, to count as independent datapoints.

Comment author: CronoDAS 14 September 2010 02:00:00AM *  5 points [-]

What about the hypothesis that terrorism, civil war, and poverty lower the median age by increasing the adult death rate?

Comment author: NancyLebovitz 14 September 2010 06:55:54AM 3 points [-]

It would take detailed research to check on it-- if parents die, babies and small children are less likely to survive.

Now that I'm thinking about it, I've got some skepticism about the census figures. I'm not saying they're wildly off, but they've got to be less accurate than censuses in first world countries.

Comment author: multifoliaterose 13 September 2010 09:04:49PM *  1 point [-]

Do you believe that experts who think developing world aid is not effective, are more likely to be driven by ideological considerations?

No, not necessarily. I think that my remark applies equally to the quote by Shiktawi and the quote by Jeffrey Sachs.

But... you just said... Wait - you mean you believe there are humans whose opinions are not heavily skewed by selection bias and/or driven by ideological considerations which have nothing to do with the available evidence?

No, although on a given topic some people are more even-handed than others. I would recur to the "A Note on Overcorrecting Bias" section at the end of my top level post.

In short: A sudden decrease in child mortality reliably causes terrorism and civil war.

The links that you provide do not corroborate your above statement (which seems to me to be quite sweeping). Are there more relevant references that you would point to?

Comment author: PhilGoetz 13 September 2010 09:15:17PM *  2 points [-]

The links that you provide do not corroborate your above statement. Are there more relevant references that you would draw attention to?

I believe that they corroborate my statement. Most countries that have a sudden decrease in child mortality, have frequent terrorism and/or civil war. Few countries today that did not have that sudden decrease, have frequent terrorism and/or civil war. Find me another variable that correlates as strongly with terrorism and war, or an argument that causation runs the other way, if you want to refute that. I find it hard to believe that terrorism and civil war cause a sudden decrease in child mortality.

Comment author: multifoliaterose 13 September 2010 09:21:42PM *  1 point [-]

I slightly edited my comment while you were replying to the parenthetical "(which seems to me to be quite sweeping)".

You raise a question which I will have to think about and may eventually have more to say something about.

My first reaction is that the collection of examples that you cite is too small to support the claim that in general to support the claim that there's even reliable correlation between decrease in child mortality to result in frequent terrorism and/or civil war.

Comment author: PhilGoetz 13 September 2010 09:36:52PM *  4 points [-]

The collection I cited is all the countries in the world today. I don't think you can get a bigger sample.

I haven't done a statistical test; I haven't searched for alternate hypotheses. The word "reliably" is too strong, since there are exceptions - I'll change it to "usually".

Comment author: CronoDAS 14 September 2010 02:02:08AM *  2 points [-]

The collection I cited is all the countries in the world today. I don't think you can get a bigger sample.

What about historically? The set of nations that exists now is a subset of all the nations and empires that ever existed. What was the median age in France in 1700?

Comment author: billswift 13 September 2010 02:32:47AM *  4 points [-]

Since de-colonization, Africa has gone from a roughly late-19th/early-20th century Euro-American equivalency to, except in specific areas, roughly a mid-19th century level of development. If you think aid does not actively harm development, what is your explanation? Do you think Africans must be inherently inferior to Europeans? Our ancestors developed, rather than regressed, despite having to invent all the systems and technology for ourselves; the Africans, and other less developed countries, know what is possible and can even fairly easily find out how to do it. So if aid is not a problem, especially if you claim aid helps, you have even more to explain about the Third World incompetence.

Comment author: multifoliaterose 13 September 2010 03:46:50AM *  9 points [-]

Since de-colonization, Africa has gone from a roughly late-19th century Euro-American equivalency to, except in specific areas, roughly mid-19th century equivalent. If you think aid does not actively harm development, what is your explanation?

So, first of all, I don't know enough about the topic to know whether the claim that you're making is correct. I would appreciate a reference supporting your claim.

A natural explanation for the alleged phenomenon that you allude is that colonization introduced foreign elements to Africa which worked okay in juxtaposition with the colonial occupation but which caused serious problems once the colonial powers pulled out on account of these foreign elements meshing poorly with the native cultures.

Comment author: billswift 13 September 2010 11:46:58AM *  1 point [-]

It isn't so much a claim as my opinion based on general reading rather than a specific reference. There are actually many potential differences and problems about development in African nations, I was pointing out that anyone making the claim that aid has any beneficial effect has to explain the apparent retrogression. For a good introduction to the problems of development, William Easterly's books, The Elusive Quest for Growth: Economists' Adventures and Misadventures in the Tropics and The White Man's Burden: Why the West's Efforts to Aid the Rest Have Done So Much Ill and So Little Good, are a good and fairly readable start.

You could also make the claim that colonization was in reality a massive aid campaign to the colonized lands, and that the problem has been caused by the effective reduction in aid since de-colonization. There have been several good studies that, contra the general belief, colonies were a net drain on the colonizers, and the colonized peoples gained more economic value than was extracted. The primary beneficiaries being a few large special interests.

Comment author: Apprentice 13 September 2010 01:48:47PM *  2 points [-]

Do you think Africans must be inherently inferior to Europeans?

I don't think Africans are inferior to Europeans and I doubt anyone reading this site would subscribe to that opinion. On the other hand I've noticed that many people around here feel that it is likely that there are some cognitive differences between human population groups and that it is likely that those have a partially genetic basis (for a readable and not overly one-sided version of this argument see e.g. here).

It's possible to imagine that cognitive differences between population groups would be important to consider in any project such as improving the lives of Africans and it's possible that this factor is currently mostly left out of consideration. Maybe this is a factor in the aid-skepticism of some of our colleagues here so maybe it would do some good to address this head-on.

I'm personally agnostic on a lot of questions here and feel a bit overwhelmed by all the factors involved. My wife and I pay for the education of a child in Africa. I hope that does some good but I can't claim any certainty on the question.

Comment author: Vladimir_M 13 September 2010 04:55:56PM *  4 points [-]

Apprentice:

(for a readable and not overly one-sided version of this argument see e.g. here).

That article you linked to is horrendously bad. This guy starts talking about multiple lines of evidence for the hereditarian hypothesis, cites a prominent paper surveying these lines, and then proclaims that the best and most decisive evidence is provided by the regression-to-the-mean phenomena. Which is in reality one gigantic logical fallacy, and quite possibly the weakest and most flawed argument ever set forth by Jensen and other prominent hereditarians. He then proceeds to spin his misunderstanding of this already catastrophically bad argument into an even more elaborate web of fallacies.

If you're interested in this topic, here are some links to recent LW discussions where you'll find a bunch of much better references. The first one specifically deals with regression to the mean:
http://lesswrong.com/lw/2nz/less_wrong_open_thread_september_2010/2jpi http://lesswrong.com/lw/2eu/open_thread_july_2010/28v5

Comment author: Apprentice 13 September 2010 08:07:55PM 3 points [-]

I've now read both threads. While I did learn some interesting things I'm not really much closer to understanding the flaws in the reasoning of the article I originally linked to.

Neuroskeptic's post on regression to the mean explains the concept clearly enough but I don't really understand why he's so hostile to it. Why does he think it's not a useful concept? The trick is being able to spot it when it happens, and to avoid being mislead by it. If you're not careful, it can happen anywhere. Uh, yeah. So why not have a word for it? Why not teach undergraduates about it? I don't get it. Did he just have a bad teacher or something?

I understand you spent a significant amount of time digging into the literature on group variation in cognition - and ultimately were unable to tell who's right. That's a bit disappointing to me, I'd hoped that with a bit of dilettanteish reading I could have an informed opinion - and you're probably better equipped to understand the literature than I am.

Comment author: Douglas_Knight 13 September 2010 09:39:53PM 3 points [-]

The term "regression to the mean" suggests time series and agency. But it's just about what happens with two correlated variables, if the correlation isn't 1. I suggest the alternative name: "a stupid prediction," as in "Because the parents' height is a stupid prediction for the children, the children of tall parents are not as tall as their parents." Perhaps it would be less vulgar to use "naive."

That's just a suggestion. Perhaps someone can think of a better name, but I think putting the focus on prediction is the key to a better name.

Comment author: liberalbiorealist 13 September 2010 09:55:27PM 0 points [-]

While you may find appeals to arguments from the regression to the mean to be "horrendously bad", I can only report that, so far as I have been able to make out, the logical legitimacy of such arguments is pretty much taken for granted among the disputants on both sides of the IQ nature/nurture controversy.

The first link you point to, which seems most directly to address the issue of regression to the mean, in turn points to papers which were written about 30 years ago or more, without, it seems, anyone in the dispute taking them seriously.

Don't you think that that would suggest that there's something deficient in the argument that use of regression to the mean in this context is a logical fallacy?

Here's the basic problem with claiming that regression to the mean in the context of, say, human traits is simply some mathematical artifact: it does nothing to explain WHY there should be a regression to the mean.

Yes, not only do the average IQs (or heights) of children regress to the mean from the average IQs of their parents; the opposite is also true -- the average IQs (or heights) of parents regress to the mean from the average IQs of their children. Does that mean that there is no causal relation established by regression to the mean effects? No, absolutely not. It only establishes that the direction of a causal arrow can't be determined from the fact of regression to the mean alone. But we know the direction of that arrow, if the cause is genetic (or environmental, presumably): it goes from parents to children, not the other way around. When we understand this, we can also explain why we see regression to the mean in the other direction as well; the same underlying set of causes are working, though, again, the direction of the causal arrow is opposite.

The fact of regression to the mean strongly argues that there is SOME underlying causal mechanism (be it genetic or environmental or a combination) that explains that fact. Why is it that the children of high IQ parents regress partly to the mean, but not all the way?

Regression to the mean in traits in both directions, from children to parents and vice versa, can be explained by luck -- those parents or children who have greater IQs or greater heights are, on average, luckier than average; they are, in particular, luckier than their own children or parents, respectively. But what are they luckier AT? What have they received more of? If one says, genes that increase the trait in question, then a perfectly coherent explanation emerges. One might say that they've received a better environment -- but that becomes a very difficult explanation in the case of IQ, since typically quite the opposite seems to be true (parents with high IQs have on average greater incomes and generally should establish a better environment for their children than they themselves experienced.)

In short, the existence of regression to the mean in the expression of traits across generations presents an important fact -- one that one might not a priori expect. Something must explain that fact. Do you seriously think that that explanatory problem simply goes away by declaring that appeals to regression to the mean constitute a "logical fallacy"?

Comment author: Vladimir_M 14 September 2010 03:50:23AM *  6 points [-]

liberalbiorealist:

While you may find appeals to arguments from the regression to the mean to be "horrendously bad", I can only report that, so far as I have been able to make out, the logical legitimacy of such arguments is pretty much taken for granted among the disputants on both sides of the IQ nature/nurture controversy.

The first link you point to, which seems most directly to address the issue of regression to the mean, in turn points to papers which were written about 30 years ago or more, without, it seems, anyone in the dispute taking them seriously.

I pointed out these papers because among the literature I’ve read on the topic, they present the best discussions and explanations of this issue. They are definitely not the last thing that’s ever been written on the subject. And while Mackenzie’s paper is indeed (yet undeservedly!) forgotten and obscure, Furby’s has been cited widely throughout the last four decades (just google for its title).

Furthermore, the logical validity of the regression argument is by no means “taken for granted” on both sides. I recommend that you read James Flynn’s 1980 book Race, IQ, and Jensen (dated, but still well worth reading), which presents a refutation of it by a prominent participant in the controversy. (It’s on pages 64-67 -- you might be able to find it on Google Books preview.) Another refutation, written by Nathan Brody, can be found in the 2003 volume The Scientific Study of General Intelligence: a Tribute to Arthur Jensen, edited by Helmut Nyborg (pages 404-407). The regression argument has also been dismissed as invalid in numerous books and papers by Richard Nisbett and many others, with refutations of varying detail and quality.

Also, an interesting critical discussion of the quality of Jensen’s statistics in general, which also addressed the regression arguments, was featured in the fall 2001 issue of the journal Chance. (Jensen himself also contributed.)

On the whole, unfortunately, a rather stupid situation has persisted since the seventies on this issue. Jensen and the other hereditarians stubbornly keep insisting on the same decades-old regression arguments, and their critics reply with more or less the same refutations. Neither side has made any further advance. However, while the anti-hereditarians can be blamed only for not coming up with more readable, clear, and in-depth counter-arguments, the hereditarians are, in my view, much more to blame because they keep bringing up the same invalid argument over and over.

(I have to add that on the whole, I have a lot of respect for Jensen as an intellectual figure, and I’m puzzled by his behavior when it comes to this particular issue. I should also stress that here I’m stating my opinion only on the specific issue of regression-based arguments, not about any other disputes that are relevant for this controversy.)

Here's the basic problem with claiming that regression to the mean in the context of, say, human traits is simply some mathematical artifact: it does nothing to explain WHY there should be a regression to the mean. [...] In short, the existence of regression to the mean in the expression of traits across generations presents an important fact -- one that one might not a priori expect. Something must explain that fact. Do you seriously think that that explanatory problem simply goes away by declaring that appeals to regression to the mean constitute a "logical fallacy"?

Honestly, with all due respect, I think you lack the necessary knowledge of statistics to reason about this issue correctly. Regression to the mean is not some unusual phenomenon that calls for a special explanation when observed. On the contrary, it is a mathematical necessity that happens whenever you have two imperfectly correlated variables (under some very generous mathematical assumptions, to be precise). For a rudimentary intuitive view, see the already discussed article by Neuroskeptic, and for detailed explanations, check out the above cited references.

In your post, you take the hopelessly muddled argument from Rushton and Jensen’s 2005 paper -- which is, incidentally, restated in their 2009 rebuttal of Nisbett’s subsequent criticism of it, thus completing another round of the decades long non-debate I described above. You then proceed to make an even bigger muddle out of it. If you insist, I can post a more detailed criticism, but if you intend to debate these topics publicly, I would advise you to acquire a greater familiarity with the relevant literature and the pertinent topics in statistics. Reading through the above listed references should give you an idea of where the problems with your argument are.

Comment author: HughRistik 14 September 2010 04:29:11AM *  0 points [-]

If you insist, I can post a more detailed criticism

As a fence-sitter on this topic, I'd put in a vote for such a discussion, when your schedule allows.

Comment author: Apprentice 13 September 2010 05:36:13PM 0 points [-]

Thank you, I will take a look at those discussions. I'm not an expert on this topic by any means and can't claim to have developed an independent opinion of this regression-to-the-mean argument. A major reason this looked to me like a convenient article to link to was that it has a collection of links to recent articles by Jensen, Nisbett etc.

Comment author: taw 22 September 2010 09:49:07PM 1 point [-]

Thank you for providing a great example that you can get a lot of upvotes on LW for totally making shit up, as long as it resonates with some LWer's prejudice.

For people naively upvoting billswift's bullshit, . For example in the most populous African country Nigeria life expectancy in 1950 was 35 years. Last year it was 48 years. You'll see such improvement in nearly every country, and this accelerated drastically in early 1990s, about the same time as Americans and Soviets stopped having proxy wars in which locals got killed.

Comment author: satt 22 September 2010 10:54:23PM 5 points [-]

For anyone else who, like me, saw taw's callout bubble up in Recent Comments and wanted to find life expectancy data to check it, a graph:

world life expectancies by continent

The bottom line (in teal) represents sub-Saharan Africa. Vertical axis is life expectancy in years, horizontal axis is year.

Comment author: taw 23 September 2010 05:10:29AM 6 points [-]

This graph seems to offer a point against my claim that progress has been recently accelerating (while totally confirming my claim that billswift is making stuff up), so I'll explain.

The late 90s drop you can see on the graph is a statistical artifact related to AIDS, early 2000s cutoff, and clustering countries into unnatural categories.

The worst affected countries were mostly in relatively well off countries like South Africa, Botswana, Namibia, Zimbabwe (it was well off until recently) etc. that don't really cluster with most of the poor Sub-Saharan Africa, and have incomes and standards of living more like Brazil or Turkey. Most of them had clear and very rapid rebounds in life expectancy since the worst in early 2000s in any case, trend-line won't be significantly affected by that.

For the poor everyone else of Sub-Saharan Africa, AIDS had much less severe effects, and for many like Nigeria I mentioned you wouldn't even notice anything just looking at life expectancy statistics. Here's per-country graphs going to 2008.

Disregarding AIDS, the most recent decade has been the best one ever for Africa by nearly any metric, be it average income, equality, poverty rates, democracy, peace, etc.

Comment author: komponisto 22 September 2010 10:16:12PM *  3 points [-]

An experiment someone should do at some point (if it hasn't already been done):

Divide (educated Western) subjects into two groups. Ask the first group how much they think life has improved in a particular African country since decolonization, without mentioning the date of the latter. Ask the second group how much they think life in that country has improved since 1950 (or whatever the actual appropriate date for "decolonization" is for the country selected), without mentioning colonization. Compare the results.

Prediction: the second group's answer will be higher.

Comment author: PhilGoetz 13 September 2010 08:15:03PM 0 points [-]

There is no incentive for anyone in Africa to become an inventor or scientist, and little incentive to become an engineer, because of all that technology outside Africa.

Of course, this is a post-hoc just-so story.

Comment author: erratio 13 September 2010 04:56:38AM 0 points [-]

One argument I have heard (no idea of the source, I might be able to google it up at some point) is that Western technology and prosperity has been achieved through the systematic exploitation of Third Worlders. This would argue that without slaves and colonies of their own to exploit, it will take Africans a much longer time to achieve the same goals.

Comment author: Konkvistador 13 September 2010 01:34:17PM *  13 points [-]

There are plenty of counterexamples. Ignoring several other ones, let me present:

Japan.

Isolationist for a good part of its history, it did attempt to conquer outside holdings before this period (they failed however). The Japanese developed a sophisticated civilization, and once exposed to Western technology and institutions they quickly caught up. They only invaded Korea after they had already reached the level of middle tier Western powers.

This doesn't prove prove that Europe didn't benefit from colonialism, what it however does prove is that its possible to develop quite rapidly without it.

Comment author: gwern 13 September 2010 01:48:32PM *  1 point [-]

Nitpick: while Japan's (first) invasion of Korea failed, Japan did take Sakhalin, & Okinawa and the Ryuku Islands early on; depending on what time periods you are considering, Taiwan might also count.

Comment author: Konkvistador 13 September 2010 04:28:00PM *  3 points [-]

All civilizations in the history of the world conquered some of their neighbours at some point. Since the earliest conquests are primarily people very similar to the ones doing the conquering and since they spend so much time together after the event, some of these events are later called unifications. Often the reason behind the similarities is the influence of previously existing empires! (the unification of Italy, the various unifications of China ect.)

What is an outside holding, the term I used in my previous post, is of course open for debate or rather definition since this is not a binary matter. However let me emphasise several ways the Japanese early expansion that you mention differ from say British expansion in Africa or India.

With the exception of Taiwan (even that region today barley musters 1/6 of Japans population) the regions are demographically irrelevant, for the purpose of "exploiting" masses of cheap foreign labour. Also even if they weren't, this wouldn't be a point in favour of the theory erratio presents since they are also all part of the first world today. The Ryuku language to top it off is very similar to Japanese having branched off a little after the 7th century. French control of Corsica or perhaps Spanish control of the Balearic Islands is perhaps comparable on the outside/inside scale. Is this really what first comes to mind when hearing the word colonialism?

The African medieval states I'll link to in a moment expanded far more in terms of sheer square km, I'm also willing to bet that they had a greater proportion of "conquered peoples" compared to Japan (the only time where this may not be true is during the height of Japanese occupation of China, but by then Japan was a developed nation):

http://en.wikipedia.org/wiki/Mali_Empire http://en.wikipedia.org/wiki/Songhai_Empire http://en.wikipedia.org/wiki/Ethiopian_Empire

I will however admit that the expansion into Sakhalin is comparable to British settlements in North America or perhaps Australia, being mostly a process of settling and taking new land with little use for the natives beyond small scale trade.

Comment author: prase 13 September 2010 09:07:12AM 9 points [-]

This is quite common argument, nevertheless I have never seen any detailed explanation of how having colonies could speed the developement. Comparison between different European countries suggests otherwise. For long time possessor of largest colonial empire, Spain, has briefly profited from colonial wealth in 16th century, a period followed by relatively slow developement leaving the country among the poorest in Europe in 19th century. There are many rich countries that never had any colonies.

And there is also quite standard argument that access to cheap labour or even slavery does more harm than good, because there is no motivation for inventions.

Comment author: multifoliaterose 13 September 2010 03:16:07PM *  1 point [-]

And there is also quite standard argument that access to cheap labour or even slavery does more harm than good, because there is no motivation for inventions.

Right, this is in line with what Collier says about having an abundance of lucrative natural resources paradoxically being detrimental to a developing world country.

Comment author: NancyLebovitz 13 September 2010 11:07:39AM 1 point [-]

No motivation for inventions, or no motivation to improve the tools that slaves use?

Comment author: prase 13 September 2010 02:49:29PM 1 point [-]

Precisely, less motivation to improve work efficiency.

Comment author: DanielLC 13 September 2010 02:10:00AM *  4 points [-]

I figure it doesn't matter if aid in general helps. It's about the aid we personally do. If we're good at finding a charity, it will be a lot better than the norm. Can anyone come up with a reason Schistosomiasis Control Initiative could possibly do more harm than good?

It matters if government aid helps when we're voting, but given the low expected difference voting makes, I don't think it matters that much.

Comment author: Douglas_Knight 13 September 2010 01:12:00PM 4 points [-]

Can anyone come up with a reason Schistosomiasis Control Initiative could possibly do more harm than good?

Did you read the link in the original post? The charity is not terribly transparent. Yes, it probably is doing net good, but there are other charities we can be more sure of that. Unless you have reason to believe that schistosomiasis is much more cost-effective to treat than other diseases, it is probably best to go with ones transparent charities with cost-effective treatments for other diseases.

Comment author: knb 13 September 2010 09:59:28PM 3 points [-]

The evidence so far suggests that as countries develop, their birthrates fall. So if aid actually increases the pace of development it should decrease the risk of Malthusian problems.

In fact, if you think granting aid to poor countries will lead to population growth, then you must be assuming resource limitations are the force that is placing a bound on population growth. If that is the case, then they're already living under Malthusian conditions anyway.

In any case, I feel pretty confident that what Africans really need is not zero sum aid but political stability and open markets to facilitate real economic growth. Botswana has had decades of rapid economic growth and it went from being one of the poorest countries in the world to being fairly wealthy. GDP per capita in Botswana is almost 40 times higher than it is in Zimbabwe. Inequality is just slightly higher in Botswana than Zimbabwe. Botswana has had property rights and largely free markets since independence, Zimbabwe, which was much wealthier before independence, has had decades of misrule.

The best thing to do would be to convince African leaders to govern like Botswana, but they have little reason to do so.

Comment author: Douglas_Knight 13 September 2010 12:49:58AM 2 points [-]

You keep linking to Holden's email as evidence against malthusian concerns, but Rosling's talk moved his opinion towards the concerns!

My main concern about people with malthusian concerns is that they never spell out what those concerns are. Maybe fatal acute famines are better than chronic malnutrition, but largely for reasons that support medical interventions.

Comment author: multifoliaterose 13 September 2010 01:58:41AM *  0 points [-]

You keep linking to Holden's email as evidence against malthusian concerns, but Rosling's talk moved his opinion towards the concerns!

Well, I guess I see Holden's email as an example of a reasonably well-informed and impartial person's take on Malthusian concerns about developing world health aid:

  1. The mainstream view is that improving health probably decreases population growth rather than increasing population growth.

  2. This view seems to be based on looking at countries in which health care has improved "organically" rather than through outside intervention and so may be unrepresentative, in particular in absence of strong women's rights & family planning the drop in population growth corresponding to improved health care may not happen.

  3. Despite the previous point, at present there's not enough evidence to count population growth as a negative unintended consequence of improving health.

My main concern about people with malthusian concerns is that they never spell out what those concerns are. Maybe fatal acute famines are better than chronic malnutrition, but largely for reasons that support medical interventions.

Can you say more about this? I'm curious. As I said, though I presently think that donating to VillageReach/StopTB has strong positive expected value I think that Malthusian concerns may be relevant and would appreciate more information.

Comment author: Douglas_Knight 13 September 2010 02:15:30AM 1 point [-]

If you have malthusian concerns, why don't you spell them out? What will happen, why would that be bad? I'm sick of arguing against amorphous concerns.

Comment author: prase 13 September 2010 10:57:50AM 3 points [-]

When population of a country like Ethiopia exceeds the level that can be sustained by their agricultural production, and still the population growth continues fuelled by the international aid, it is a problem for distinct reasons:

  • The country will be dependent on aid. A policy change on the side of the donators could cause famine, and the country's government could be easily blackmailed and become a puppet of their sponsors.
  • We aren't successful in helping African countries to develop more efficient economy today. After their population doubles, it would be even more difficult.
  • There is certainly some limit which the population couldn't surpass even with the aid of aid. When the population reaches that limit, we can expect famine, war or massive emmigration. Those would happen sooner without the aid, however in smaller scale, because the population would be smaller.

One can hope, of course, that somehow Africa will become developed and the population will stop growing enough soon. But there is little evidence that this will happen. The kind of aid which supports population growth doesn't solve the problem of poverty. It only postpones the solution while the problem is growing in scale.

Comment author: whpearson 13 September 2010 02:37:59AM 2 points [-]

I'm unsure of the exact claims.

I expect one of them might be that any life's saved might be lost in future rwanda-style genocides. If that was exactly malthusian.

I'd worry about encouraging population growth in water scarce areas as well. It would give less time for technical measures to be developed to meet the water needs.

Comment author: Douglas_Knight 13 September 2010 04:38:22AM *  3 points [-]

If the infant is saved from diarrhea and dies 20 years later in a genocide, I'd say that is good for the infant. Perhaps one should be careful about calling that 20 DALYs, rather than one life. Of course, increased population may make war more likely. If one is a total utilitarian, the repugnant conclusion seems to me like a slam-dunk here. If one is an average utilitarian, this looks at first like a poor trade (let's average just over a country, not the world).

Treating diarrhea probably has the effect of reducing not just mortality, but morbidity. When we saved the infant, we treated 100 other infants who would have lived. But now they'll have healthier lives, probably happier and more productive. I think that should satisfy the average utilitarian. They'll be more productive if all things are equal, but the increased expectation of catastrophe--famine or war--decreases perceived life-expectancy, and thus productivity. War is worse than famine in that it destroys property as well as people. I'm not sure where the balance lies, but I think productivity is improved by choosing war over diarrhea.

Productivity is important because it seems linked to demographic transition. Maybe malthusian concerns say that the best interventions are those which are best connected demographic transition, like girls' education. But as to the question of whether curing diarrhea is a net benefit, I think it quite likely.

Amartya Sen is quite enthusiastic in choosing acute famine over chronic malnutrition (and China over India more generally). Disease is quite like chronic malnutrition in its effects on health and productivity. The more serious malthusian complaint is if we end up with chronic malnutrition. If we're trading disease for chronic malnutrition, I don't think we're making things worse, though it may not be worth the bother. Solving acute famines to leave chronic ones is probably making things worse.

Comment author: multifoliaterose 13 September 2010 03:37:58AM 1 point [-]

Fertility rates are very high in Africa as indicated by this graphic. If saving lives in the developing world does not indirectly result in a drop in fertility then it pushes in the direction of promoting exponential population growth with a large base.

In general I'm worried about the possibility of global natural resource shortage corresponding to our high rate of natural resource usage resulting in a future natural resource shortage which leads to a population crash, political instability of the nuclear powers, permanent obstruction of future technological development or even human extinction.

I'm worried that pushing in the direction of faster population growth may result in a global natural resource shortage sooner rather than later, before we have the chance to develop sustainable solutions for the demands that human lifestyles impose on Earth's limited resources.

Comment author: Douglas_Knight 13 September 2010 05:37:31AM 2 points [-]

There is so little trade with Africa that its increasing population is not going to affect the global resource market.

Comment author: Konkvistador 13 September 2010 01:43:25PM *  6 points [-]

You do realize that African populations migrate to Europe and the Middle East by the hundreds of thousands per year and that by current projections the rate will soon rise to millions?

Comment author: multifoliaterose 13 September 2010 06:26:11AM 1 point [-]

I upvoted your remark because this seems like the start of a useful conversation which has the potential to resolve my concerns.

What you say is true at present. But what about taking the long view, projecting ahead to Africa's hypothetical technological development? The question is whether Africa will develop before or after we've developed sustainable solutions. I would guess that we'll be noticeably past peak oil by the time Africa develops, but not completely sure about this, and maybe there are other future natural resource shortages to look out for?

Comment author: khafra 13 September 2010 12:44:11PM 2 points [-]

There's a sizeable minority that believes we're at peak everything. I'm 75% certain aid to Africa will be at less than 10% of its current levels within two decades, either because we've basically dealt with the problem of resource scarcity, or because the developed world needs every scrap of fuel, food, and materiel it can hang on to.