Month: November 2005

Paying for Performance II

Roland Fryer’s experiment to pay school children for better grades will go into effect next year reports the New York Post.

Under the pilot, a
national testing firm will devise a series of reading and math exams to
be given to students at intervals throughout the school year.

will earn the cash equivalent to a quarter of their total score – $20
for scoring 80 percent, for instance – and an additional monetary
reward for improving their grades on subsequent tests….

said details about the number of exams, what grades would be tested,
funding for the initiative – which would be paid for with private
donations – and how the cash will be distributed are still being
hammered out.

"There are people who are
worried about giving kids extra incentives for something that they
should intrinsically be able to do," Fryer said. "I understand that,
but there is a huge achievement gap in this country, and we have to be

Thanks to Katie Newmark for the pointer.

Why are rental cars American cars?

Our new hire at GMU, Ilia Rainer, posed this question over lunch yesterday.  Why don’t rental car companies use the superior Japanese product?  Our group came up with a few hypotheses:

1. Rental car drivers consume patriotism by using the American product.  Often a third party, such as a corporation, is picking up the bill.  Rental car companies don’t want a "foreign" image.

2. Rental cars are (and must be) well taken care of by Hertz and others.  Japanese cars perform better when maintenance is low, but with plenty of care American cars do just fine.

3. Here is a variant on #2: Rental cars have higher value on the resale market than regular used cars, given that they are well taken care of.  This boosts the value of U.S. cars relative to Japanese cars, since Japanese cars will hold up anyway.

4. The fraud problem in the auto repair market is severe.  If you can fix your cars yourself, at marginal cost, U.S. autos are a fine buy.

5. U.S. cars are more comfortable for long drives, which makes them better suited for the rental market.  They are also better for "nature driving" out west.  Japanese cars are better for daily commutes, urban driving, and stop-and-go driving, which are more likely found in your daily life and less likely relevant for the rental market.

But surely you can do better.  Your thoughts?  Comments are open.

In case you missed it

Here is the latest on the Steve Levitt controversies concerning abortion and crime.

Here is a Crooked Timber symposium on Jonathan Strange and Mr. Norrell.

Here is Jane Galt on home electronics.

Here is Daniel Drezner on whether macroeconomists are more interventionist than microeconomists.

Here is Arnold Kling on Tyler Cowen’s Law.

Here is Andrew Gelman on Bryan Caplan, I love the line about pessimistic bias.

Pledge Drive

It’s that time of year again, our NPR moment.  Do you enjoy Marginal Revolution’s blend of economic commentary, book reviews, art recommendations and fun ideas?  Please consider making a donation.  You can donate through Amazon or via the PayPal link in the upper left hand corner.  Any amount would be appreciated.  Use the PayPal link for donations over $50.  We are also happy to accept donations via snail mail.  Send to:

Marginal Revolution, c/o Alex Tabarrok
Department of Economics, MSN 1D3
George Mason Universtiy
Fairfax, VA, 22030

For donations of $100 or more we would be pleased to send you an autographed copy of one of our books, your choice (including my forthcoming book Judge and Jury available in February).

If you are in a position to be especially generous and wish to make a larger contribution through a legally recognized non-profit entity (for tax reasons),
please let us know, this can be arranged.

Why are we asking for donations?  MR is never going to be a paying venture but donations help us
to cover our costs.  More importantly, donations help to solve a serious economic problem.  Efficiency says that goods with zero marginal cost should have a zero price but without prices not only is the incentive to produce diminished but so is information about what to produce.  (See Coase’s 1946 classic, The Marginal Cost Controversy, JSTOR).  Donations allow prices to be set at MC while at the same time providing a (noisy) signal about where true economic value lies.  In particular, Tyler and I know that we can appropriate more of our marginal product from professional work than we can from blogging yet it is conceivable that our marginal product is higher in blogging.  Thus, to decide how much to invest in this venture we markup donations to get an estimate of our social value and we put positive weight on social welfare in our utility function.   

Happy Holidays!

A Humean thought experiment

Let us say, just for fun, that you woke up one morning to a world where everyone else’s demand curve — except yours — slopes upward.  But it is not common knowledge that  this is the case.  What is the first oddity you would notice? 

1. The most expensive radio stations would be filled with the most ads.  The music would never come.

2. Your house would have no electricity, due to grid overload.  (Is this true?  An upward-sloping curve does not mean you will demand more at the current price.  Think of twisting the demand curve around the current point of intersection.)

3. The most transparent agents would be found wandering the streets, bereft of all wealth, the victims of corporate price hikes.

4. You would wonder why so few people were reading your blog.

5. You would check ebay and find very high prices for items with active bidding.  (Hmm…what kind of auction markets are behind the scenes for our power supply?)  (Addendum: What is the Nash equilibrium here?  Will people hold off bidding, hoping that tomorrow’s price will be higher?)

6. You would be puzzled why the Nordstrom’s sale was so empty.

7. It would take you days to notice any significant difference at all.  After a few weeks, they would call in the econometricians to solve the identification problem in the data.

How long would it take you to figure out that other peoples’ demand curves were sloping upwards?  How long would it take for society to fall apart?

Comments are open.


FolderShare is a very cool service that synchronizes folders in real-time on two or more computers.  I can work on Stata files at the office, for example, and by the time I get home the same files will be on my home computer.   No more forgetting to shuttle the latest update of my work from office to home or vice-versa!  FolderShare is also useful for transfering large files to a co-author.  You can open folders on your computer to a guest, for example, and let them synchronize files up to 2 gigabytes in size.  No more mailing of CDs!  And oh yes, it’s free!


Rudy Rummel writes that due to new evidence he has significantly updated his figures for 20th century democide,  i.e. murder by government.

Many scholars and
commentators have referenced my total of 174,000,000 for the democide
(genocide and mass murder) of the last century. I’m now trying to get
word out that I’ve had to make a major revision in my total due to two
books. One is Wild Swans: Three Daughters of China by Jung Chang, and the
other is Mao: the Unknown Story that she wrote with her husband, Jon
Halliday. I’m now convinced that that Stalin exceeded Hitler in
monstrous evil, and Mao beat out Stalin.

From the time I wrote my book on China’s Bloody Century, I have held to these democide totals for Mao:

Civil War-Sino-Japanese War 1923-1949 = 3,466,000 murdered
Rule over China (PRC) 1949-1987 = 35,236,000 murdered

some other scholars and researchers had put the PRC total as from
60,000,000 to a high 70,000,000. Asked why my total is so low by
comparison, I’ve responded that I did not include the China’s Great
Famine 1958-1961. From my study of what was written on this in English,
I believed that:

(1) the famine was due to the Great Leap Forward when Mao tried to catch up with the West in producing iron and steel;
the factorization of agriculture, forcing virtually all peasants to
give up their land, livestock, tools, and homes to live in regimented

(3) the exuberant over reporting of agricultural
production by commune and district managers for fear of the
consequences of not meeting their quotas;

(4) the consequent belief
of high communist officials that excess food was being produced and
could be exported without starving the peasants;

(5) but, reports from traveling high officials indicated that peasants might be starving in certain localities;
(6) an investigative team was sent out from Beijing, and reported back that there was mass starvation;
(7) and then the CCP stopped exporting food and began to import what was needed to stop the famine.

I believed that Mao’s policies were responsible for the famine, but he
was mislead about it, and finally when he found out, he stopped it and
changed his policies. Therefore, I argued, this was not a democide.
Others, however, have so counted it, but I thought this was a sloppy
application of the concepts of mass murder, genocide, or politicide
(virtually no one used the concept of democide). They were right and I
was wrong.

From the biography of Mao, which I trust (for those
who might question it, look at the hundreds of interviews Chang and
Halliday conducted with communist cadre and former high officials, and
the extensive bibliography) I can now say that yes, Mao’s policies
caused the famine. He knew about it from the beginning. He didn’t care!
Literally. And he tried to take more food from the people to pay for
his lust for international power, but was overruled by a meeting of
7,000 top Communist Party members.

So, the famine was
intentional. What was its human cost? I had estimated that 27,000,000
Chinese starved to death or died from associated diseases. Others
estimated the toll to be as high as 40,000,000. Chang and Halliday put
it at 38,000,000, and given their sources, I will accept that.

I have to change all the world democide totals that populate my
websites, blogs, and publications. The total for the communist democide
before and after Mao took over the mainland is thus 3,446,000 +
35,226,000 + 38,000,000 = 76,692,000, or to round off, 77,000,000
murdered. This is now in line with the 65 million toll estimated for
China in the Black Book of Communism, and Chang and Halliday’s estimate
of "well over 70 million."

This exceeds the 61,911,000 murdered by the Soviet Union 1917-1987, with Hitler far behind at 20,946,000 wiped out 1933-1945.

perspective on Mao’s most bloody rule, all wars 1900-1987 cost in
combat dead 34,021,000 — including WWI and II, Vietnam, Korea, and the
Mexican and Russian Revolutions. Mao alone murdered over twice as many
as were killed in combat in all these wars.

Now, my overall
totals for world democide 1900-1999 must also be changed. I have
estimated it to be 174,000,000 murdered, of which communist regimes
murdered about 148,000,000. Also, compare this to combat dead.
Communists overall have murdered four times those killed in combat,
while globally the democide toll was over six times that number.

Scream this from the rooftops

Ed Glaeser writes in his new abstract:

Does bounded rationality make paternalism more attractive? This Essay argues that errors will be larger when suppliers have stronger incentives or lower costs of persuasion and when consumers have weaker incentives to learn the truth. These comparative statics suggest that bounded rationality will often increase the costs of government decisionmaking relative to private decisionmaking, because consumers have better incentives to overcome errors than government decisionmakers, consumers have stronger incentives to choose well when they are purchasing than when they are voting and it is more costly to change the beliefs of millions of consumers than a handful of bureaucrats. As such, recognizing the limits of human cognition may strengthen the case for limited government.

Yes, some NBER Working Paper Abstracts should be screamed from the rooftops.  Here is the paper itself.  Glaeser also offers some arguments against "soft paternalism":

1. Soft paternalism is an emotional tax on behavior which yields no government revenues.

2. Soft paternalism can cause bad decisions just as easily as hard paternalism.

3. Public monitoring of soft paternalism is much more difficult than public monitoring of hard paternalism.

4. While hard paternalism will be limited by public opposition, soft paternalism is particularly attractive because it builds public support.

5. Soft paternalism can build dislike or even hatred of subgroups of the population.

6. Soft paternalism leads to hard paternalism.

7. Soft paternalism complements other government persuasion. 

Get this:

Soft paternalism requires a government bureaucracy that is skilled in manipulating beliefs.  A persuasive government bureaucracy is inherently dangerous because that apparatus can be used in contexts far away from the initial paternalistic domain.  Political leaders have a number of goals, only some of which relate to improving individual well-being.  Investing in the tools of persuasion enables the government to change perceptions of many things, not only the behavior in question.  There is great potential for abuse.

Thanks to Daniel Klein for the pointer.

Practice questions for my macro class

The students themselves are to write the questions into the comments section.  Do it soon, and yes that means you.  Then the students should practice these questions in their spare time, with the clock ticking.  Ideally each question, or at least some of them, should come as a surprise.  Don’t read them all until you are ready to give them a try.

Most people study to make themselves feel better about doing their work, and not to actually succeed in their chosen field of study.  They spend hours staring blankly at sheets of paper.  They should spend more time trying to solve problems or answer questions, usually under simulated exam conditions and with a clock ticking. 

Tick, tick, tick…That’s the way to go, and yes I know it hurts.

Burkina Faso fact of the day

Over 2001 and 2002, America’s 25,000 cotton farmers received more subsidies — about $3 billion — than the entire economic output of Burkina Faso in a year.  Two million people in Burkina Faso live partly or fully from cotton farming.

The information is from Raising Less Corn, More Hell: The Case for the Independent Farm and Against Industrial Food, by George Pyle.  The book is more libertarian and less anti-corporate than the title makes it sound.

A free lunch for Medicare?

At present, Medicare, as the system is called, cannot be claimed abroad.  So American pensioners tend to travel back to the United States to get treatment.  The possibility of making Medicare "portable" has been talked about for years.  But, apart from the introduction of a small pilot project, it has never got much further than just an idea.  Yet the advantages are clear: expatriate pensioners would find it easier to get health care; the costs for the crisis-ridden Medicare would be lessened; and Mexico and other Central American countries with American pensioners would benefit not only from a rise in their health-care expenditure, but also from the big increase in their numbers that such a change would certainly bring.

That is from The Economist, "Go South, Old Man," 26 November issue.  Comments are open, in case you know more about this.  Here is more information, which includes a brief discussion of fraudulent claims.