Fractal Wrongness:

Fractal Rightness::

Photos from this morning's walk down to and along the Singapore river

Institutional Change

"What's the point of knowing the world if you cannot change it?"

Theories of institutional change:

• functional/efficiency theories

  North 1993: When groups see opportunities that cannot be realized under existing institutions, they make new ones. (Examples: new companies. bankers inventing new financial instruments. Property rights). North's ideas are tricky because they are so obvious that it's hard to tell if they are genius (things which are obvious to all but only after a genius figured them out) obvious or tautologically obvious: people create institutions because they need new institutions in order to get stuff they want.

• public interest
• power
• path dependence
• evolutionary
• Bayesian Learning/demonstration effect
• Crisis as trigger

Exam notes

35% of final grade. Computer lab 2. From 12 problems, choose six. For each problem, describe in one paragraph and look at appropriate institutional mechanisms. E.g., poverty, rent-seeking, regulatory capture, externalities. Think about enforcement, about principles behind mechanism. Discuss hypothetical consequences of adopting the prescribed institution: solve problem, don't solve problem, perverse incentives.

Final Paper

Due May 5.

Student Presentations

I've generally given up the struggle to comment productively and politely about student presentations, in favor of the uncontroversial point that the LKY school should put more resources into training students on public speaking and presentations at the beginning of the school year, and should do something to get professors to be a bit more rigorous and consistent on evaluating student presentations. And I do have one more point:

A simple test to determine if a recommendation is meaningful or just bullshit is to see if offers a real choice. Would you actually consider doing the opposite? Consider:

• get on the "learning curve"
• build capacities

Is that helpful? Did you depend on that advice to help you NOT do the following things:

• get off the "learning curve"
• destroy capacities

That aside, I want to make a point about cancer. There's a serious problem in how cancer statistics are interpreted by scientists, doctors, the media, and the public. I think Gird Gigerenzer's book was my first exposure to this paradox. It's this: cancer screening is not a purely positive thing, and actually may be a bad thing in some cases. Let's look at how this could be true:

Take a simple country with 1000 people, who live to age 70, and only one kind of cancer. If there is no cancer screening or cancer treatment, 10 will die of cancer, and all at age 60. Since there's no cancer screening, these ten cases are not discovered until they have severe symptoms, let's say at age 58. So the average survival duration after cancer detection is 2 years.

Now restart our clocks and add cancer screening, every two years starting at age 40. This time, 10 cases of cancer are discovered, all at age 54. Everybody gets treatment. They all die at age 60. The average survival duration after detection has risen to 6 years. But all that really happened was that ten people each spent an extra four years dealing with cancer. They didn't actually have longer, better lives.

Now, let's go one more time around, adding super-sensitive cancer screening. This time, 20 cases of cancer are discovered, all at age 50. Everybody gets treatment. Many go into remission, but ten still die of cancer at age 60. The rest die of other causes at age 70. The average survival duration after detection has risen from 2 to fifteen years! But in fact, nobody lived any longer than they would have without screening, and twenty people lived as cancer survivors for years or decades, having paid in money and blood and tears for treatment that didn't actually help.

Think this model is absurd?

AN Australian researcher says there's little evidence that prostate cancer screening saves men's lives.

And Professor Simon Chapman of the University of Sydney said a study of Australian newspaper and television stories about such screening for prostate cancer found most of them promoted it aggressively, ignoring the almost complete absence of evidence that it would save lives. — The Australian

The point is that not all cancers will kill you, at least not before something else will. We can detect cancers that we can't effectively treat, and we can't always differentiate between cancers that will kill you and cancers that won't. And it's a fallacy to say it's always better to be safe than sorry, because it doesn't work that way. False positive results, being told you have cancer when in fact you aren't slated to die from cancer, can lead to more than a little sorrow, especially if you undergo expensive and painful unnecessary treatment. The National Cancer Institute in the US says the same thing, but in a much more convoluted way:

At least two requirements must be met for screening to be efficacious:

1. A test or procedure must be available to detect cancers earlier than if the cancer were detected as a result of the development of symptoms.
2. Evidence must be available that treatment initiated earlier as a consequence of screening results in an improved outcome.

These requirements are necessary but not sufficient to prove the efficacy of screening, which requires a decrease in cause-specific mortality. For example, these two criteria are met in the case of screening for childhood neuroblastoma by assessment of urinary catecholamine metabolites. On the basis of these criteria, a mass screening program was conducted in Saitama Prefecture, Japan, between 1981 and 1992 for 6-month-old infants.[3] Over that 12-year period, the annual incidence of neuroblastoma in children younger than 1 year increased from about 28 per million to 260 per million but without a significant reduction in incidence in children older than 1 year. Because there also was no reduction in mortality for the disease, this experience provided strong evidence of overdiagnosis—diagnosis of some neuroblastomas detectable by screening, which would not have been clinically diagnosed later. Similar experiences have been reported elsewhere in Japan [4] and in the Quebec Neuroblastoma Screening Project (QNSP) in Canada.[5] The history of screening for neuroblastoma also provides a useful illustration of the benefit of undertaking well-designed evaluations of emerging screening technologies before implementing screening programs. Although such studies are very costly, it has been shown that the QNSP itself averted unnecessary morbidity for thousands of children and did so while returning a yield plausibly estimated at a cost savings 64.5 times the investment in the study.[6] —NCI

In case you didn't follow, let me translate:

Cancer screening is a bad idea unless there's a test that finds cancers early, and treating these cancers early actually helps. Even then, screening may not be a good idea. They did a twelve-year test in Japan where they found way more brain cancer in infants under age 1, but cancer detection rates in older children didn't change and on average nobody lived any longer. So screening infants for brain cancer (at least, with that kind of screening and that kind of brain cancer) was a big waste. That 64.5 times savings they mention is, if you read carefully (and I had to check the abstract of footnote six to be sure I had it right), is the savings from scrapping unnecessary cancer screening programs, not the savings from performing screening. Bury the lede much?

Remember, these are general points. This is not a diatribe against all screening, or in favor of cancer. But it is clear that screening is not an unmitigated positive, and it's a big mistake to think it is. This is a tough point to make in the face of powerful individual appeals from survivors, but the underlying issue is the same as other kinds of medicine: individual testimonials are not data. If you have a mental picture of someone dying unnecessarily from a late diagnosis, you need to balance it with a mental picture of someone dying unnecessarily from treatment for a cancer that they don't actually have. Then you can put all of this emotion to the side and get back to evidence-based medicine. Put another way, humans are not wired to think accurately about statistics, and we need to remind ourselves of this weakness constantly.

Today's good news:

Spain's re-elected Prime Minister Jose Luis Rodriguez Zapatero announced a government Saturday which for the first time included more women than men and a female defence minister. —AP

Meanwhile in Singapore:

Prime Minister Lee Hsien Loong reshuffled his Cabinet on Saturday, as part of a process to groom the next generation of leaders for Singapore.

The most eye-catching in the slew of changes announced by PM Lee was the appointment of Mr K. Shanmugam as Law Minister and Second Minister for Home Affairs from May 1. —Straits Times

The newspaper featured pictures of twenty faces: from the PM and Minister Mentor and Senior Minister to the various Ministers. As many as five looked Indian or had Indian names; at least one had a Muslim name. None appeared Malay, and none were women.

This is both a symptom and a problem. It's a problem because it restricts the viewpoints present in discussion by Singapore's ruling group (though, arguably, not any more than one-party rule). Some Singaporeans brag about the meritocratic government recruitment system, but it's a poor excuse for meritocracy that doesn't raise up any women or many minorities.

The equivalent for the US: five women out of twenty one. Four ethnic minorities out of twenty one. For most cabinet members I've never heard anything about their religion, so I would assume all are professed Christians.

Side note: Did one second of Google on that assumption, and found the most adorable Hegelian dialog within the Radio Islam web site (that's Farrakhan):

• Jews in Bush’s Cabinet? Don’t Hold Your Breath
• USA´s Rulers -They are all Jews !

It's not quite as tasty as it looks, since the first link is a reprint from the New York Observer and the second refers to Clinton's administration. But it did lead me here, which (taken with, obviously, a very large grain of salt) suggests that there are no Jews in Bush's cabinet. I'm sure there are no Muslims, Hindus, or Buddhists. I can just barely imagine one of them being a (very discreet) atheist, but I haven't heard anything. Of course I wouldn't be the first to suggest that many of them appear to actually practice Manichaeism. This is getting a bit off-topic and I'm using too many fancy words, so let's just re-iterate the positive: a majority female Spanish cabinet, and female heads of government in Argentina, Chile, Germany, and Liberia. This post should not be construed as a declaration of partisanship in the US Democratic Presidential Primary race of 2005-2008.

Service Delivery for the poor

Per-capita spending on health doesn't correlate with better health conditions for poor. One key reason is that, in almost all countries, most of that spending is concentrated with the wealthy, so per-capita spending on the poor remains very low.

Guest Speaker

Sasa Vucinic from Media Development Loan Fund. See him speak here.

We start with this video.

Only 18% of the world's population lives in countries with a free press. "Only 18% knows exactly what's going on, gets unfiltered views ..." Um, there went a big non sequiter? Are we going to talk about media failings in countries with freedom of the press?

Got financing from Soros in 1996. Financed 167 different projects worldwide with 70 teams in 23 countries. Current loan portfolio $23m with interest rate 2% to 10%, lowest rate to riskiest projects. Overall, 2.5% loss rate. One case of fraud, in Macedonia, involving imported chicken wings.

Funding is from large agencies, however they have inconsistent direction due to management turnover. Now offering bonds for individual investors.

How not to censor a radio station: send troops to confiscate the equipment. Very photogenic. How to censor: call the potential advertisers and threaten them.

Student presentations.

CPF changes in 2007

CPF is Singaporean Social Security. The first presenter spends 6:30 giving a recitation of minor details of the 2007 changes to the CPF, presented devoid of context or interpretation. Example: "By legislating re-employment by 2012, to require employers to offer re-employment to workers reaching 2 up to age 65, and eventually to 67".

Second presenter: Analysis. Will the reforms encourage people to work longer? will they benefit low/middle income/older? Will the resource be enough [sic]? More reading verbatim from slide.

Do the changes actually benefit low/middle? Not really.

Questions: What does it mean that the plan allows reduction of older workers' wages?

Joel's evaluation:

• Reading/reciting text from slides: check.
• Poor time management: check
• Slides filled with hundreds and hundreds of words in small text: check
• Overall thoughts: might be a good presentation to a group of Singaporean accountants. As a presentation of policy analysis, dry and difficult. Dives directly into excessively detailed exploration of minutiae without without any introduction, explanation of context.

Prof's evaluation: presentation doesn't refer to any economic concepts from class.

Public Housing Subsidy: UK vs Singapore

• Reading/reciting text from slides or notes: check. Softer-spoken reader, too, so the soporific effect is amplified. I want (as I do with most student presentations) to go up and rip the notes out of her hands and shout "snap out of it!" Next presenter in the group much better on this account. Update: no, I take that back. As she's gotten deeper into her time period, she's talking way too fast as she simply shovels out words in a stream. If you have 20 minutes, you need to reduce the number of ideas to fit, not just talk faster. And you need to present a hierarchical structure: here are my main ideas. Now I'm talking about the first main idea. Now I'm telling you three details in support of my first main idea. Now I'm telling you my second main idea. Etc. Instead, this is just a flood of words that goes on and on, almost in a chant, without any use of pauses or any other structure, so once you tune out, you're gone.
• Poor time management: No. finished 2 minutes early
• Slides filled with hundreds and hundreds of words in small text: check
• Overall thoughts: Once again, the presenters' focus is on the minute facts and details that they have researched, not on introducing an idea to the audience. These presentations are way, way, way too detailed in the wrong, wrong wrong places. All facts and data, little to no information or knowledge.

Singapore Healthcare Policies and Financing

In 2005 Singapore spent 3.8% of GDP (with government contributing a total of 0.9% of GDP, or about a quarter.) Would love to see that directly contrasted with other countries (the US as the most expensive, Japan, the other Tigers, etc). Oh, now we get some pointless small details.

• Reading/reciting text from slides or notes: Check. Just as painful as the last one. If you are reading your entire presentation verbatim from a piece of paper, you are failing. If you can't hold the ideas of your presentation in your head, how do you expect your audience to do so? Second presenter is much better on these terms, though he's picked up the hyperfast talking problem. (I also talk way too fast when presenting.)
• Poor time management: finished 4 minutes early. borderline between good and too short, but certainly far better than too long.
• Slides filled with hundreds and hundreds of words in small text: check. Bonus demerits for using clip art figures with incongruously pink skin. Further demerits for pointless animation. Credit for actually stating a conclusion more or less clearly, but still the same absurdly busy, pointlessly overformatted, essentially randomly organized slides as everybody else.
• Overall thoughts:

This (PDF) is the carbon cap and trade game that I created when it was my group's turn to present in Global Issues and Institutions class. It has two intended pedagogical purposes: to demonstrate how a carbon market will move production to more efficient producers (which is the the point of such a market), and to illustrate how cheating might undermine the benefits of such a market. I playtested it once, and then we played it once in class, and I discovered a third pedagogical purpose: to illustrate that smart, educated people may still have trouble creating a rational market in the context of a confusing, complicated new game :).

I created the game because a quick search didn't show anything quite like it. What I specifically wanted to create was a carbon emission permit market that could be created in a classroom without special equipment. This isn't a general-purpose "learn about global warming" game, it's an attempt to model a real, if very simple, permit market. It clearly needs more work, so I'm now offering the materials of the game on the internet, under a CC Attribution license (do whatever you want with it, but include my name and link back to this post).

Here's what didn't work when we tried it. (In some of these, "work" is subjective because we wanted to see a particular outcome for educational reasons; so there's a catch-22 between rigging the game to get the result you want, and setting up the game as an experimental simulation and seeing what happens. In general I think the model is so simple and contrived that it doesn't have any experimental information to offer; it should be treated purely as a teaching tool):

1. We didn't come anywhere near playing eight or twelve rounds in 30 minutes. We did two or three rounds in 1-2 hours.
2. The results over the few rounds we played weren't as expected. Overall efficiency should have improved, but it got worse. This probably reflects individual players not clearly understanding what actions to take. (Note that this also depends a huge assumption in the game, that companies that make more profit per unit of emission also make more economic output per unit of emission.)
3. Almost nobody cheated. Admirable but not realistic.
4. The paper bits for money and permits are a big pain in the ass. I think it would go better computerized, so that everybody's actions are forced to follow the rules of the game and so that we wouldn't need to do any data collection during the game. Of course, then you would need a computer lab or laptops to play.
5. I thought I rigged all of the companies so there would be some obvious buyers and obvious sellers, but that happened only shakily.
6. The instructions don't help make a working market. This is really the big problem underlying the others, because this is the crux of the home game. The game needs some clear notes on how to run a real auction market with buyers and sellers, so that everybody who wants to buy or sell can do so easily in a few minutes.

So, here it is. If you want help using it in your class, I'd be happy to help you.

The instructions PDF for the game can be downloaded here. The spreadsheet supporting the game is available here, in OpenOffice Calc format.

This work is licensed under a Creative Commons Attribution 3.0 United States License.

The Ionian Mission, Patrick O'Brian

Inspired by Naomi Novik's novels, I'm now working through O'Brian's twenty Aubrey and Maturin books. They are amazing. I don't plan to review any of them any further than that. I do want to offer this passage from The Ionian Mission, the eighth book:

Now when the fiddle sang at all it sang alone: but since Stephen's departure he had rarely been in a mood for music and in any case the partita that he was now engaged upon, one of the manuscript works that he had bought in London, grew more and more strange the deeper he went into it. The opening movements were full of technical difficulties and he doubted he would ever be able to do them anything like justice, but it was the great chaconne which followed that really disturbed him. On the face of it the statements made in the beginning were clear enough: their closely-argued variations, though complex, could certainly be followed with full acceptation, and there were not particularly hard to play; yet at one point, after a curiously insistent repetition of the second theme, the rhythm changed and with it the whole logic of the discourse. There was something dangerous about what followed, something not unlike the edge of madness or at least of a nightmare; and although Jack recognized that the whole sonata and particularly the chaconne was a most impressive composition he felt that if he were to go on playing it with all his heart it might lead him to very strange regions indeed.

From the body of a Reuters article:

... the survey, conducted among 69 U.S. rock-formatted stations in markets as diverse as Los Angeles and Knoxville to Buffalo, found 84 percent of the respondents planned to vote in the November election.

...

About 30 percent of the respondents called themselves Democrats, while nearly 22 percent described their politics as Republican and 21 percent declared they were independents.

...

Asked about their overall presidential preference, Democrat Barack Obama led the pack with 26 percent support among those planning on voting in the November election. McCain ranked second with 22 percent and Democratic candidate Hillary Clinton drew 18 percent.

The survey found women and fans of alternative radio, featuring '80s and '90s rock, tend to be Democratic, while men and classic rockers lean Republican.

So Democrats are a plurality of respondents and Obama is the leading candidate. What's the headline? "Male rock fans likely to vote Republican: survey"

I realized over the weekend, the same weekend that included 16 hours of Leadership and Communication training between from 2:30 Friday afternoon and 5 pm Saturday, that I've mentally checked out for the semester. This is not the fault of the training program, which was taught as well as possible for such a large class at such an inconvenient time. (In fact, that kind of public speaking training, as well as basic presentation skills, ought to be moved up to the pre-semester 1 special classes.) Nonetheless, my brain has departed for greener pastures.

View Larger Map

Unfortunately, it's only week 11 out of about 13, not to mention reading week and two weeks of exams. While my attendance is unblemished (so far) and I've written or transcribed or quoted about eighty thousand words of class and lecture notes over the last three months, the semester is far from finished, and my groupmates in my three or four or five group assignments would probably not appreciate my permanent mental departure. So, I guess I'll just mentally check back in and get back to work.

The World Bank

Note that there's a very substantial difference between the World Bank and the IMF. The World Bank has a lot to do with development, whereas the IMF is a bunch of economics PhDs.

Is the World Bank part of the problem or part of the solution? (Or both?)

Joel's note: When I started thinking about getting a post-graduation job, I talked to some professors and started reflecting on my options. One of my epiphanies was that, even though these multinational agencies and NGOs and such have global scope and deal with issues all over the world, they all have offices. If you work for one, you'll probably work in one of its offices. And those officies are in specific cities. If you want to work for one, you'll probably need to live in its city. Like these, in Washington D.C:

View Larger Map

In 2006 the World Bank approved US$23.6 billion in loans. Global FDI was US$1.3 trillion, and total capitol flow was $6 trillion. (The US economy was $13 trillion that year.)

The World Bank IDA offers 40-year plus interest-free loans, in theory funded by repayments from older loans but in fact always short of funds. Accounted for $9.5 billion of the World Bank's total 2006 lending. "In contrast to previous studies we find that the US exerted a significant influence on IDA lending during the period 1993 - 2000. We demonstrate that the influence was both statistically as well as economically significant." (University of Copenhagen study)

You can't talk about the World Bank without mentioning the Washington Consensus. And here's a useful-looking Stiglitz interview:

Multinational Monitor: What was the "Washington Consensus?"
Joseph Stiglitz: It was a consensus formulated between 15th Street and 19th Street in Washington among members of the International Monetary Fund (IMF), the U.S. Treasury Department, and the World Bank. It argued that the keys to success in developing countries were three things: macro-stability, liberalization (lowering tariff barriers and market deregulation) and privatization. It was largely formulated out of experience with Latin America.

MM: Does it still exist as a consensus?
Stiglitz: No. There is a consensus that those precepts, while important, are neither necessary nor sufficient for successful development.

Institutional Economics is a large part of the post-Washington Consensus thinking.