Saturday, July 23, 2011

Terrorism, Extremism and other bad stuff

Update(07/25): evidence is mounting that Breivik was part of a larger group and that one of his objective was to create a revolution. That something I would qualify as terrorism...

When I somewhat disagree with two of the most interesting people on the intertubes, Glenn Greenwald and Juan Cole, I feel really sad inside. So let me expose my sadness.

Norway's double-disaster yesterday led to some comparison between the treatment of an attack by a Muslim and by a Blond Norwegian. On this, I completely agree with the first part of Glenn Greenwald's post(until the part on the definition of terrorism). Linking yesterday's attacks to Al-Qaeda because the methods employed are the same is quite a weak link, for instance. The initial statements of some media(TV and newspapers) were quite bad given what we know now. Also, as Juan Cole says:
In 2008, only one terrorist attack out of hundreds in Europe was committed by radical Muslims. In 2010, according to Europol [pdf], 7 persons were killed in terrorist attacks. Some 160 of these attacks that year were carried out by separatists. The number launched by people of Muslim heritage? 3.

Where I am more doubtful is on the fact that yesterday's crime should be considered as terrorism, and that it is not qualified as such just because the guy is blond. The same argument was made when some guy crashed his plane on an IRS office last year.

My feeling is that those guys are crazy. They might have an extreme ideology. But it is not clear that they are trying to instill terror. One of Al-Qaeda's specific objective is to create fear. After Osama Bin Laden's death, there was discussion on whether he has won or not, because of the cost of the Iraq and Afghanistan war, the increasing security costs(e.g. passenger screening). They pride themselves in low-cost actions leading to high-cost reactions:
In his October 2004 address to the American people, bin Laden noted that the 9/11 attacks cost al Qaeda only a fraction of the damage inflicted upon the United States. "Al Qaeda spent $500,000 on the event," he said, "while America in the incident and its aftermath lost -- according to the lowest estimates -- more than $500 billion, meaning that every dollar of al Qaeda defeated a million dollars."
Calling an act a "terrorist" act implies that the act was not the goal in itself, but that the goal was to instill fear that it could be repeated. That is why the plane crash in the IRS building is hard to qualify as "terrorism", for instance.

At FirstPost, a security specialist at NYU in London asks the key question about the Norwegian bomber/shooter:
"The next key question is whether he was acting alone, or whether he is part of a group."
Indeed, the group thing makes the act repeatable, and makes the act part of a larger objective. Before calling the act a terrorist act, this information is key. Now, today, Blake Houndshell tweettussed a document that surfaced on the internet with a blurry origin. The document seems to be some kind of manifesto, which would make the act part of a more global plan. That would be terrorism.

In the end, that's probably a question of terrorism. But I feel that before we know the true motivation for yesterday's disaster, saying that the media is bad because they don't call the guy a terrorist is not a good point.

The difference between polls and forecasts

Update(07/25):  " I realized that polls and forecasts are close. They take coarse signals, and try to derive the "state of the world'' from them." says basically nothing. Sorry

I am going to write a couple of posts on polls and forecasts, because this is one of the topics I am working on right now. Basically, I want to understand what information  we can get through polls. The motivation was quite simple:

  • Polls are used and have predictive power(Nate Silver relied on polls for the 2008 election)
  • Polls are often wrong(The French election of 2002 did not exactly go as expected. The Dewey/Truman contest in 1948 did not go as expected either)
The conclusion of that was that polls are partially informative, and it could be interesting to try to understand how much information is actually conveyed. So this led me to think about polls. Then I realized that polls and forecasts are close. They take coarse signals, and try to derive the "state of the world'' from them.

Let me give just two quick examples of what I mean by polls and forecasts. Polls are all the number you see that tells you the share of the vote Obama get in 2012 or how many people want to raise the debt ceiling. Forecasts try to predict what GDP growth will be in 4 months.

Polls and forecasts provide various incentives. The initial discussion on rational expectations equilibrium, by Grunberg and Modigliani(1954), started with the question on whether there exists an accurate forecast. The issue is similar to the Heisenberg principle: once the poll or the forecast is out, it changes the information set. This means that even if it was "accurate" in the first place, the very act of its revelation will probably make it inaccurate. If the polls are saying that the National Front is at 30%, some people might consider turning out at the polls while they would have stayed home otherwise.

Is that a problem? My feeling is that it depends on the goals of polls and forecasts.  Polls and forecasts seem to differ in their objective.

In theory, we want the poll to provide accurate information, in my opinion. The reasons polls and forecasts are made is in order to provide information. In forecasts, this seems obvious. Decisions are based on forecasts, and we can assume that the better the information, the better the decision. One issue here is that if a forecast changes decisions, it might change some assumptions on which the forecast is based, thus making the forecast inaccurate, and the decision inefficient. Therefore, the forecast is more about rational expectations: it has to take into account the impact it will have on the very assumptions it is based on.

In polls, things looks different. Why do we have polls in the first place? Polls measure the state of the world at a moment in time. From these, decisions are made. For instance, political candidates might change their platform. The decision to take a vote on an issue might change. A company might change its motto or its product. This means that the poll should provide the ex-ante accurate information, and should not care about its impact on what it polled. Agents will react to the poll, and in a perfect and infinite world, we could probably assume that this back and forth between agents and polls will end up in an equilibrium. The issue is that this infinite world does not exist: we do not have an infinite time for polling, and the agents are not infinitely rational.

Therefore, forecasts aim at being accurate ex-post while the polls only aim at being accurate ex-ante. Those are the criteria on which they should be evaluated.

Why is this relevant? I started thinking about rating agencies, given the fuss about their behavior in the European debt crisis and the raise of the US debt ceiling. Here is a primer. Rating agencies are not a real problem. The incentives are bad, of course, when they are paid by the agent they rate. However, the main issue is that their rating are taken into account by law. For instance, one big issue of the last couple of weeks was that the ECB wouldn't accept Greek bonds as collateral if Greece was considered in default. Some pension funds can only invest in tripe A bonds, from what I understand. 

This is the real problem. Even with the negative outlook on US debt by Moody's and S&P, the interest rate on US Treasuries did not move. If there are no rules to force them to invest depending on ratings, investors might not care about the rating agencies, if they think their rating is not just. My gut feeling is that rating agencies are "polls", and should be treated as such: they provide information at a date in time, and agents might respond to it. But there should be no strict constraint in how people act on this information.

Internet changes how we remember

A recent study published last week tried to show that Internet changes the way we memorize. I do not know why this did not get more attention, because I find that incredibly important. This might be biased by the fact that I have a terrible memory and that Google comes in handy, when, for instance, I need numbers to make a point. The main point of the paper is that
[W]hen people expect to have future access to information, they have lower rates of recall of the information itself and enhanced recall instead for where to access it 

The interesting general process at work is transactive memory. Usually, this has been studied in couples. For instance, there is a nice study from Wegner et al.(1991) which looked at dating vs. non-dating pairs and found two interesting results:

  • Dating pairs were better at memorizing things than non-dating pairss, when nothing was asked by the experimenters
  • The results were reversed when the experimenters imposed a "structure", i.e. told the members of each pair what type of information each should remember

The traditional conclusion is that couples unconsciously separate which member of the couple should remember which type of information

The last experiment of the paper had a really interesting results. Participants could save information to a folder on a computer. They were then tested to see whether they remembered more where the information was than what the information was. 50% of the participants recalled the folder where the information was stored. 23% recalled the information. 17% recalled both the information and the folder, and amazingly, only 11% recalled the information but not the folder.

Another part of the paper mentioned at the beginning is somewhat related to an earlier post where I mentioned my preference for continuous examination rather than a one-day big exam, with the French baccalaureat as an example. One experiment from the study was quite relevant. The experiment asked participants to remember answers to questions. The questions were asked, such as "An ostrich’s eye is bigger than its brain". The participants could then look up the answer on a computer. The treatment was that

Half the participants believed the computer would save what was typed; half believed the item would be erased.(...) 
They find that those who believed the information would be erased had a far higher rate of recall than those who thought it would be saved. The general idea for this result has been known as directed forgetting, which is the term you want to use to show-off when someone explains to you that they've learnt stuff for their exam with the objective of unlearning it when it becomes unnecessary. Interestingly, knowing that they would be tested on it did not matter for the participants:
Participants were more impacted by the cue that information would or would not be available to them later, regardless of whether they thought they would be tested on it.

I reproduce the cartoon of the earlier post...

Disclaimer: One of the co-author is an assistant professor of psychology at Columbia University

Saturday, July 16, 2011

Washington's tardiness - Follow-up

Related to my last post, here is Stan Collender, former House and Senate budget committee staffer, on the perception he had of the debt ceiling debate back in February when he met a "Tea Party" meeting at the House of Representatives:
What I took from that was, first, that the debt ceiling was going to be a lot more trouble than anyone realized. They did not want a negotiation there. There was a religious-like fervor on that point: Voting for the debt ceiling was a sin, and you can’t just sin a little. Second, Boehner and Cantor were going to have a lot more trouble than anyone thought.

Thursday, July 14, 2011

Washington's tardiness

As a quick note, it's important to remember that since 2003, everybody knew that the Bush tax cuts would expire in 2010, but the debate dragged big time during the lameduck session last year. And this is also happening with the debt ceiling. Rand Paul gave an interview to Christiane Amanpour the week after the 2010 midterm, and said that he would vote against raising the debt ceiling. When Amanpour tells him that the US would then default, Paul argues that this would only be the case if the Republicans actually won the vote, implicitly saying that he thought that his position was a minority one. What he wanted was, in the pure tradition of Washington protest speaking, "to send a message to Washington"(to pronounce with the accent).
That is to say that there is a frustrating shortsightedness among US politicians which is quite annoying.

On Taxes - Follow-up

To continue on my last post, Dylan Matthews directs us to Piketty and Saez's paper on historical US tax rates between 1960 and 2001. The first interesting figure is the cross country comparison of the marginal tax rate for the different income percentiles. Those rates have declined everywhere, with the decline far more pronounced in the US

The second interesting graph shows the composition of tax rates in 2004 and in 1960, with a huge fall in corporate and estate taxes. It is to note that 2004 comes after the two rounds of the Bush tax cuts(2001,2003)

 Finally, the inequality in the distribution of incomes has risen, with a sharp break in 1986.
In the paper, Piketty and Saez provide evidence that the causation does not run from lower taxes to rising income shares, which is interesting in itself.

Wednesday, July 13, 2011

Good diagnosis and bad arguments on taxes

Ezra Klein refers to Martin Wolf's astonishment at the tax debate in the US:

The astonishing feature of the federal fiscal position is that revenues are forecast to be a mere 14.4 per cent of GDP in 2011, far below their postwar average of close to 18 per cent. Individual income tax is forecast to be a mere 6.3 per cent of GDP in 2011. This non-American cannot understand what the fuss is about: in 1988, at the end of Ronald Reagan’s term, receipts were 18.2 per cent of GDP. Tax revenue has to rise substantially if the deficit is to close.

This is quite misleading. One big story of the recession is that tax revenues dropped drastically. It's hard to compare this year's share of tax revenues to "the Reagan years" or to the post-war average. Jared Bernstein has something similar:

 Tax revenues are at their lowest as a share of GDP, but once again, the economy still is far below capacity...One better comparison is with tax revenue and government spending in the last recession, in 2001. Via...Ezra Klein:
 In 2001, revenues were at 19.5 percent of gross domestic product and spending was at 18.6 percent of GDP. That was our surplus. In 2010, revenues were at 14.9 percent of GDP while spending was at 23.9 percent.
The right comparison would actually be to consider the revenues if taxes were held at today's level(what the Republicans want), in a "normal economy". So one possibility is to look at historical tax rates and see if today's are far below historical levels(and be careful of the loopholes).  You can also compare the rates with other countries
 For instance, the top marginal tax rate
Jared Bernstein had a nice comparison or corporate income tax rates between the US and other countries, and over time
Because I am on my old computer (and so it takes hours to connect to a website and find pictures), I'll stop there for now, but a little bit of googling will show you that the tax rates are usually at historical lows. This means that Ezra, Jared and Martin are right, but the tax revenue number as share of GDP is, in my opinion, the bad way to frame the argument.

Tuesday, July 12, 2011

Prevention and addiction, with tipping points

In the Tipping Point, that I mentioned a couple of times in previous posts, Malcolm Gladwell has a Freakonomics-like approach to different societal problems. That is to say, we should probably take what he says with pinches of salt. This being said, he mentions two things about prevention and addiction that are quite interesting.
The first thing that caught my attention was the study of Leventhal and Singer(1965) who looked at how they could better convince university students to get a tetanus vaccine. The students were given either a "high-fear" booklet or a "low-fear" booklet, that could be put in parallel with a low-fear cigarette pack and a FDA-approved cigarette pack. The students in the "high-fear" treatment were significantly more convinced to go take the shot. However, when the authors actually considered who took the shot, there was no difference between the two groups.
They also provided some of the students with a map giving the directions to the health-center where to get the shot, and they got a 28% increase in the number of students getting the vaccine. Interestingly, once again, there was no difference between the high-fear and low-fear treatment in the response to the "map-treatment".

That's all nice for prevention, but what about addiction? The second part that I found interesting was focused on smoking(yep, that's original). There were two things there:

  • First, Gladwell mentions studies showing that you get a high success rate in making people quit when you treat smokers for depression. Apparently, the biological reason comes as follows: "Bupropion(an antidepressant) does two things. I t increases your dopamine, so smokers don't have the desire to smoke, then it replaces some of the norepinephrine, so they don't have the agitation, the withdrawal symptoms".
  • Second, Gladwell argues that nicotine addiction is not linear, though he might be a bit biased here since with linearity, you would have no Tipping Point. Basically, he argues that every person has an individual threshold(different for everybody) in the number of cigarettes they can smoke before being addicted. Though he's not mentioning it clearly, it seems that that would lead him to advocate a really gradual reduction in the level of nicotine in cigarettes: you need to make the change non-perceptible, so you need graduation, but the more you lower it, the less addicted people you'll get.

Lords of Finance - On Bankers

There is an interesting parallel on the perception of bankers after 1932 and the one today. As Ahmed puts it "Bankers were now increasingly viewed as crooks and rogues". He then mentions the usual congressional hearings, with some fancy stories that sound familiar:
[The public learned] that Albert Wiggins, president of Chase, had sold the stock of his bank short at the height of the bubble and collected $4million in profits when it collapsed(...) Charles Mitchell, (...) of the National City Bank, had lent $2.4million to bank officers without collateral(...) only 5 percent of which was repaid(...), J.P. Morgan had not paid a cent of income taxes in the three years from 1929 to 1931.

Lords of Finance - How to efficiently transfer gold from Britain to France

After 1928, France started accumulating an increasing amount of gold while the rest of the major countries were fighting the start of the Great Depression. The gold standard was showing its weaknesses: to provide credit, bank needed enough gold reserves, but fear led investors to direct capital towards the US and France which already had the largest reserves. Interestingly, capital flows from the US to Europe dried up because of this uncertainty, but also because of the Smoot-Hawley Act of 1930 which imposed huge tariff hikes for US imports. This led to a drop in demand for European goods. In part for that reasons, European countries thus could pay their own imports mainly using gold, and same thing to pay its debt. The flows in Europe were mainly directed to France, which started with reserves of about $1bn in 1928 to triple that in 1932, while the UK, who started in 1928 with a similar amount, saw its own reserves dwindling:
The interesting tidbit in Ahmed's book is about how money was transferred from Britain to France. The reserves of the US, Britain and France were actually mostly held in London, because of the weight of bullion. Therefore:
[C]entral banks had taken to "earmarking" the metal, that is, keeping it in the same vault but simply re-registering its ownership. Thus the decline in Britain's gold reserves and their accumulation in France and the US was accomplished by a group of men descending into the vaults of the Bank of England, loading some bars of bullion onto a low wooden truck with small rubber tires, trundling them thirty feet across the room to the other wall, and offloading them, though not before attaching some white name tags indicating that the gold now belonged to the Banque de France or the Fed.

This is also, by the way, how I was transferring ownership of objects with my brother in our room when we were young.

Lords of Finance - Irving Fisher

I continue my posts on Lords of Finance(see here for instance) with an entertaining description of Irving Fisher by Liaquat Ahmed. Irving Fisher got his name from the Fisher rule(or maybe it's the opposite), the one that tells you that the nominal interest rate is equal to the real interest rate plus expected inflation, which then tells you that the real interest rate is usually negative when you're at the 0 lower bound. Apparently, Fisher was weird:
Originally a mathematician(...)Fisher was quite a[n] odd bird(...). [H]e had emerged from the sanatorium a committed vegetarian. He suffered from terrible insomnia and, to cope with it, had designed a bizarre electrical contraption that he hooked up to his bed and was convinced helped him to fall asleep. He was also a proponent of selective breeding and was secretary of the American Eugenics Society; he believed that mental illness originated from infections of the roots of the teeth and of the bowels and(...) was a fervent advocate of Prohibition

Wednesday, July 6, 2011

Delegating decisions

A couple of times in the last semester. I heard the question on why people obey the law, why populations are following the rules decided by their government and legislature. I was struck that the first and only thing that came up was the constraint story: the government has constraining power through, say, its regalian authority.

Now, there is no denying that this plays a part. However, it seems to me that one strong reason people abide by the rules is that they have elected those people to make decisions in their stead. In the same way you don't have time to investigate the news by yourself and you expect the journalists to do this job for you. This was underlined in a recent post of Dean Baker on the coverage of Minnesota's government shutdown by NPR:
It is not balanced reporting to present a Republican legislator from Minnesota talking about spiraling state spending and then present someone else talking about state services. Most NPR listeners will not have the time to look up the data on state spending in Minnesota. NPR's reporter should.
In the political field, Gabriel Lenz published a paper in 2009 where he considers the earlier evidence of a "priming" effect in political campaigns. Priming is a general psychological theory in which a prior exposure to an external signal makes you more receptive to a future signal or more prone to a certain behavior. If you want a fancy example, a recent psychology paper suggested that male libido was a main cause of war

Across four experiments Lei Chang and his team showed that pictures of attractive women or women's legs had a raft of war-relevant effects on heterosexual male participants(...)The effects on the male participants of looking at attractive women were specific to war. For example, their ability to locate pictures of farmers, as opposed to soldiers, was not enhanced. Moreover, the war-priming effects of attractive women were greater than with other potentially provocative stimuli, such as the national flag.
Previous research tended to explain the fact that the voters are giving more importance to topics discussed in news media to judge the quality of the candidates they have to vote on. In his paper, Lenz explains that there are two possible ways to explain why you might observe such a relation:

  • People might learn from the news which candidate supported their ideological position and will judge them based on this. For instance, you might have an isolationist ideology. Initially, you do not know which candidate in the Republican primary feels the same way. Watching the debate last week, you realized that there might be differences in the Republican party, namely, you realized that you want to support Mitt Romney's position
  • People might learn from the news the positions supported by each candidate and adopt the position of his favorite candidate. For instance, if you're a Mitt Romney fan and that you had no clear position on Afghanistan, you might change your mind and push for a quick withdrawal, after what he said in the Republican debate
Interestingly, Lenz finds far more evidence for the latter effect: people tend to adopt what their favorite politician/party defend. The fact that this happens might be obvious, the fact that it is by far the most important effect was not clear to me. Admittedly, Lenz's paper only focuses on 4 different case studies, but the evidence seems to be quite clear that at least on those topics, the leaders have a huge influence. It also shows that people, maybe consciously, maybe unconsciously, delegate their power of decision to the politicians.

Monday, July 4, 2011

Build a robust system, not a safe system

In a recent interview at The Economist, Tim Harford(who blogs at the Undercover Economist) sells his new book: Adapt: Why Success Always Starts With Failure

I have not read the book, but given the interview, I think this is in line with a couple of things I mentioned in earlier posts. In the talk, Harford refers to Donald Rumsfeld's request not to use the word "insurgent" and all its derivatives when referring to the war in Iraq. It's one of those examples where you build a system that might work in the short term but is prone to shocks in the long term. More generally, this is what you see with the Obama Administration trying to control the leaks of classified documents. As Geoffrey Stone puts it in a NYT editorial:
President Obama has also followed Mr. Bush in zealously applying the state secrets doctrine, a common-law principle intended to enable the government to protect national security information from disclosure in litigation.
No matter where you stands theoretically on whether the government should have the ability to classify various things as "secret", or "top secret", there is a practical reason why the current policy is counter-productive. You create a huge incentive to whistle-blowing, and a leaked information is worse than a communicated information in the impact it has on the population's perception of the information.

In France, for instance, the Bettencourt trial basically ruined Eric Woerth's career when it was revealed, among others, that his wife had been hired to manage Liliane Bettencourt's wealth. You might worry that divulging this information will get the media over you for conflict of interest, but this is a sure thing if you do not divulge the info and the media finds it through a whistle-blower or a police investigation, for the very reason that you tried to hide it.

Classifying information is risky for this reason: you build a system that might work in the short run, but a "shock"(say, some guy downloading 150000 diplomatic cables on a Lady Gaga CD), will make the system completely unsustainable.

Saturday, July 2, 2011

Lords of Finance - On the origin of the summer season on the Riviera

When Benjamin Strong and Montagu Norman, the US and British central bankers, decided to take a vacation in southern France in the summer of 1926, the southwest of France had just become a holiday destination. The two bankers stayed at the Hotel du Cap, which
Like most resort hotels on the Riviera, it used to shut between May and September. However, in 1923, a rich young American couple, the Murphys, persuaded the owner to keep it open and took over the whole hotel for the summer. Thus was born the summer season in the south of France
Well, that was easy. This seems a bit far-fetched to me, but there must be a Tipping Point story buried inside...

Lords of Finance - On the wealth of political leaders

In France, we've heard a lot about the train de vie of Dominique Strauss-Kahn, with the Porsche incident, with the conditions of his initial house arrest($250000 per month for security and other measures), or with his dinner yesterday to "celebrate"(?) the new developments in his New York trial
Dominique Strauss-Kahn, the former head of the International Monetary Fund, celebrated the lifting of his bail conditions over attempted rape allegations by enjoying a $600 meal with his wife at one of New York's smartest restaurants.
Interestingly, there were other famous political leaders who enjoyed nice things, though different cars. From Lords of Finance:
Churchill was addicted to high living He had a Rolls-Royce and a chauffeur and by his own admission had never been on a bus or on the Underground. He kept an enormous retinue of twenty-four servants, and pampered himself with the finer things of life - silk underwear, champagne at every meal, Havana cigars, strings of polo ponies, and bouts at the gaming tables of Monte Carlo and Biarritz - and was predictably in perpetual debt.
The bottom line is that people might forget your fancy lifestyle if you're competent and you're good at your job, and if you don't sleep with prostitutes.

Lords of Finance - On the origins of French aversion to speculation

I recently finished Lords of Finance, by Liaquat Ahmed. The book is quite interesting, though unnecessarily detailed, and looks at the impact of the 4 Central Bankers of Great Britain, France, Germany, and the US, on the economy during this period. The message is not that positive: the subtitle is "The Bankers Who Broke The World".

I won't make a summary, but I wanted to underline some interesting tidbits from the book. The first one is on French aversion to speculation.

The first thing that caught my eye was Ahmed's discussion on where French awareness of speculators comes from. Every now and then, Nicolas Sarkozy blames speculators for the rise of commodity prices of for the Greek crisis. Interestingly, Ahmed links this French spirit on the post-war crisis of the Franc in the early 1920's. He argues that after the war, France had to spend $4 billion on reconstruction, but refused to raise revenues under the assumption that the Germans would eventually repay their debt and contribute to the effort. Moreover, France had spend $30billion on the war effort, and because the French were(at the time), extremely tax-averse, revenues contributed for only 5% of this sum, the rest coming from borrowing from French middle classes($15billion), foreign governments, and printing money. Eventually, the amount of currency in circulation tripled during the war, while it only doubled in the UK.

This meant that after the war, the French franc lost a lot of its value against the dollar: before the war, the ratio was 5 Francs for $1, stabilized at 15 to 1 in the early 20's, but reached 20 to 1 at the beginning of 1924 because of the failure of the invasion of the Ruhr, casting pessimism on the possibility to finance the reconstruction spending through German reparations.

The negotiations over the Dawes plan, - a plan supposed to restart the European recovery, hurt by the recent bout of German hyperinflation, the subsequent default of $12billion of reparations and the occupation of the Ruhr by France and Belgium -, made the Franc fall 27 francs to the dollar on March 8th, because, as I understand it, the plan would reduce the amount of reparations the Germans would have to pay in the short term thus reducing France's ability to pay for reconstruction other than by printing money. Ahmed argues that at this point, Prime Minister Poincare and other officials started evoking a great conspiracy theory headed by the German government to attack the franc through the holding by German business houses of 13 billion francs in reserves.
The French were then, and would remain for many decades, obsessed with the specter of foreign speculators. Keynes described their attitude in the preface(...) for the french edition of the Tract on Monetary Reform: "Each time the franc loses value, the Minister of Finance is convinced that the fact arises from everything but economic causes. He attributes it to the presence of a foreigner in the neighborhood of the Bourse or to the mysterious and malignant influences of speculation"

A primer on education and reforming the baccalaureat

I've been discussing about education with a friend of mine recently, and I thought I should write a couple of things.

First, on education. My based-on-no-evidence belief is that there are two important characteristics an education system should feature:

  • Repetition: The teachings should be repeated over and over, and the exams should test on the same thing over and over. A recent study indeed showed that repeated teaching and repeated retrieval was quite awesome: 
Repeated retrieval with long intervals between each test produced a 200% improvement in long-term retention relative to repeated retrieval with no spacing between tests. However, there was no evidence that a particular relative spacing schedule (expanding, equal, or contracting) was inherently superior to another
As an example, in The Tipping Point, Malcolm Gladwell discusses the success Blue's Clues, an educational TV show for kids, and explains that "Nickelodeon runs the same episode for five straight days". In each episode, a dog finds clues to an enigma, and children usually try to find the solution as the clues come along. The show proved to be even more popular than Sesame Street
  • Involvement: A lecture does not work, a high level of interaction is necessary. Practically, it also means that I like the anglo-saxon system where lectures serve as support and a book is usually required to work by yourself on the thing you're studying. I think involvement is something that everybody mentions, but I thought I would write it here. Apparently, a large body of research has been focused on this since Astin(1984) and his theory of involvment Again, Gladwell mentions that the implication of children is primordial in their learning through the TV shows: "the child at home is given the opportunity to shout out an answer on his own. Sometimes, Steve[the host] will play dumb"
What does it mean practically? The leak of one math exercise of the French baccalaureat(an exam at the end of high school) restarted a debate on whether the "bac" was actually relevant, or if it should be replaced by, say, continuous evaluation.

My prior on this is that continuous evaluation is the first best, notably given the two priorities I mentioned above. A single examination at the end of high school is not the best way to test people's knowledge. I found this nice Calvin&Hobbes cartoon to summarize the point:

Now, what is the goal of the Bac? What we want is to test people's knowledge. As a consequence, the Bac is used as the selection criterion for universities for instance. It also means that it has to be comparable across students so that the test yields the same information for everybody.

So, the Bac gives a signal of one student's "ability" or, more generally, "type". The result is a signal to companies and universities that you are competent. Now, the problem with a signal is that it is not perfect. For instance, you could have thought that Michael Bay just made a mistake with Transformers 1, but with 2 other Transformers you know far better how good he is.  A signal based on a single evaluation is a really noisy signal. To get rid of the potential large errors on any single signal(e.g. you just eat some E-Coli infested sprouts just before your exam), you want to use multiple evaluations.

Hence, not only multiple evaluations give a better signal of one's competence, but as I mentioned above, multiple evaluations are actually better for anybody's knowledge. If we want to test people, but also make them more knowledgeable, multiple evaluations seem the way to go. That's why I think it is the first best, and that's why I think the only reason the Bac still holds is because we're afraid of touching universality and the ability to compare.

The main issue I've seen on forums and such is that if we only look at the grades given by professors to students during their high-school, we'll have problems because:
- students and families can make arrangements with professors,
- schools are not equal, and grades obtained at one school are not comparable with grades obtained at another school.

The issue here is that people define continuous examination as considering the grades that one gets at his or her school by his or her professor during the year. But continuous evaluation, for me, just means a system where the student is evaluated continuously. That is, a regular evaluation at the scale of the department or the region is also a continuous evaluation. Moreover, evaluations at this scale should get rid of the problems of student/professor arrangements, or problems inherent to particular schools. More generally, what you want is to design a system where the scale enables you to get rid of the professor/school idiosyncrasies.

In any case, I think the debate on the Bac vs. continuous evaluation in France won't go forward if we keep this narrow definition of continuous evaluation. We can and should devise a better system where students are evaluated more frequently, because that'll improve retention by students and that'll improve our evaluation of students.