Deinonychus antirrhopus: terrible claw

Friday, January 31, 2003

 
Kaus Defends Krugman...and Flunks Statistics Mickey Kaus decided people were too harsh on Prof. Krugman's last column. Kaus writes:

Matthew Dowd, Bush's pollster, put out a release noting that Bush's Gallup approval rating had dipped only three points, from 63 to 60. Donald Luskin and Andrew Sullivan oh-so-predictably piled on! But what Dowd didn't point out is that Bush's disapproval number increased seven points over the same period in the same poll, from 29 to 36 -- for a total swing of 10 points.

Problem one: You can't add the disapproval ratings to the decrease in approval ratings. It isn't like, disapproval = 1 - approval.

Since the first Gallup poll after the election (when Bush rose a bit), the total swing has been 17 points. Not a "plunge," but not chopped liver either.

Problem two: Krugman stated that the President's approval rating had plunged in the last two months (say January of 2003 and December of 2002). Going back to the November election effectively makes it 3 months. So which is it? Two months or three months?

I'd cut Krugman some slack on this one, given his imperiled state -- threatened by Bush oppo goons, his cage waiting for him in Guantanamo, operating "under much more scrutiny than any other opinion columnist," beset by pygmy bloggers, etc. ...

Sorry, but this is the wrong attitude. He gets no slack after he points out that he is bright enough to do his own analysis. He can't even present the number honestly. He could have said, "since the midterm election...", or "in the last three months....". Just because he is only less misleading than in other columns doesn't mean we should give him a pass.
Steve 10:30 AM

 
28 Pakistani al-Qaeda Arrested in Italy The police have found explosives, forged documents and maps of Naples with "sensitive areas" circled.

At least the Italians conduct routine searches for illegal immigrants. Makes you wonder who is coming through the U.S. borders. Especially in light of this news report that notes that U.S. agents with forged documents easily entered the U.S. And then there is this news report where the INS has put its estimate of illegal aliens at 8 million up from 1.5 million from its last estimate. Further, the INS now estimates the population of illegal immigrants is growing by about 350,000 a year vs. its last estimate of 275,000 a year.
Steve 9:29 AM

 
Two German's Convicted of Selling Arms to Iraq

MANNHEIM, Germany (AP) -- A German court has convicted two businessmen of breaking arms export laws by helping Iraq to acquire drills that can be used to bore a "supergun" capable of firing nuclear, biological or chemical shells.

Steve 9:13 AM

 
Turkey: Put more pressure on Iraq Could this be and indication that Turkey is coming around to the Bush Administration's position?

"If the military grip around Iraq is tightened then they will be more encouraged to agree to a peaceful solution without firing a shot," Yakis said at a news conference after talks with leaders of the European Union, which Turkey wants to join.

Steve 9:09 AM

 
Prime Minister Berlusconi's Comments at the White House

THE PRIME MINISTER: We will never forget that we owe our freedom -- our freedom -- our wealth to the United States of America. And our democracy. And we also will never forget there have been many American young lives that were lost and sacrifice themselves for us.

So for us, the United States is not only our friend, but they are the guarantee of our democracy and our freedom. And I already has the opportunity to say this to President Bush, every time I see the U.S. flag, I don't see the flag only representative of a country, but I see it as a symbol of democracy and of freedom.

Steve 9:03 AM

Thursday, January 30, 2003

 
The Cato Institutes Take on the State of the Union Speech One thing Cato notes right up front is that Bush scaled back the number of proposed programs from the previous year. Last year Bush proposed 39 new programs, but this year proposed only 20. Compared to Clinton that is a meager number. In his 2000 SOTU, Clinton proposed 104 programs! I like Cato's list under Compulsory Compassion

Compulsory Compassion
9. "I propose a 450 million dollar initiative to bring mentors to more than a million disadvantaged junior high students and children of prisoners."
10. "So tonight I propose a new 600 million dollar program to help an additional 300,000 Americans receive [drug] treatment over the next three years."
11. "[T]onight I propose the Emergency Plan for AIDS Relief.... I ask the Congress to commit 15 billion dollars over the next five years..."
12. "I urge you to pass both my faith-based initiative..."
13. "...and the Citizen Service Act."

See, the Republicans really are nice guys and not the meanies the Democrats say they are.
Steve 3:59 PM

 
Dow is Cutting 4,000 Jobs Ouch! Not good news for Bush, the economy seems to be getting weaker not stronger. And even with the accelerated tax cuts I don't see Bush's plan doing anything anytime soon.

Update: Coca-Cola is cutting 1,000 jobs.
Steve 3:52 PM

 
U.S. Senators Question Iraq Evidence Senators are asking Secretary of State Colin Powell to appear before Congress to explain what new evidence links Saddam Hussein to terrorism.

Senator Christopher Dodd said that the American people needed to hear the evidence against Iraq before the UN, and asked for Mr Powell to appear before the full Senate next week to present his case.

I have to admit I want to hear what this evidence is myself.
Steve 3:49 PM

 
Economic Anti-Idiotarianess A very good article on why liberals often have a hard time with economic logic and thinking. The basic message, they just aren't used to it.

Actually it is more involved than that and the criticism doesn't just apply to liberals (although my impression is that liberals suffer from economic idiotarianess more frequently than non-liberals). The idea is that economic logic is something rather new and has not played much if any role in evolution. Other types of thinking such as the communal forms of thinking have (hunter-gatherer socieities for example) been around for a long time and have played a role in evolution. Hence these types of logic are more easily grasped by the brain. So the typical response is:

Communal logic (I understand it therefore it is good).

Market/Economic logic (I don't understand it therefor it is bad).

You can see this with the notion that the profits from drug companies are bad or evil, the copyrighting of drug treatments for AIDS is evil or bad, etc. These people ignore the problems inherent in intellectual property (that it exhibits many of the characteristics of a public good and thus "free riding" can result in under production). That without some of these sub-optimal outcomes we might not even have the drugs at all.

Anyhow check out the article it's pretty good. Hell it even has some Krugman bashing at the end (yes Krugman with a Ph.D. in economics uses economic idiotarian rhetoric).

(Via Instapundit)
Steve 11:07 AM

 
Shooting Down the More Guns, Less Crime Hypothesis Well there I am getting back, finally, to the Ayres and Donohue paper. I started in on the empirical section of the paper and the authors make reference to figures and tables. I look through the rest of the 55 page document and no tables or figures. Hmmmm...maybe they are in a seperate file, sure enough 46 pages of figures and tables are in a seperate PDF file. Ugghhh, this paper practically doubled in size! Oh well.

My frist problem is Ayres and Donohue's discussion of figure 1a which looks at Robbery. In looking at the figure the robbery rates (defined as robberies per 100,000 population) one sees that the robbery rates tend to go up and down based on 5 to 10 year periods. Now, the authors break the data down into four different trends. The first is the 22 states that never adopt shall issue CCW laws, the 8 states that adopt a shall issue CCW law post-1977 and pre-1990 and 17 states that adopt shall issue laws post-1989 and then 4 states that are considered to have a shall issue law pre-1977. Now, the first problem I have here is that aggregated data (which these are) can hide information. It is possible that one state in one of the series is driving things or distorting the aggregate trend. It would have been nice if the authors checked for this and reported what they found.

Another problem related to figure 1a is the tentative conclusion that Ayres and Donohue speculate on. The make the following comment on top of page 12

Indeed, if one were forced into making causal attributions from this graphical data, one might conclude that shall-issue laws increase robbery rates.

Their reasoning is that because in the 22 never adopting states the robbery rates are driven down so far at the end of the time period that they are the same as the eight states that adopted shall issue laws pre-1989 (that is relatively speaking the robbery rates in those 8 states has increased compared to the 22 states). The problem is that Ayres and Donohue know that one should not make such an inference even if forced to. On page 11 Ayres and Donohue make write

The second point that leaps out from the figure is that the 22 states that have not adopted shall-issue laws have had much higher rates of robbery than states that allow the carrying of concealed handguns, at least until recently (more about this later). Note that this is not what Lott and Mustard mean when they suggest more guns, less crime. They realize, as sophisticated researchers, that in 1977, the 22 never-adopting states had double the robbery rate of the other states for reasons having nothing to do with their lack of shall-issue laws. Indeed, only four of the other 29 states allowed the carrying of concealed handguns at that time.

In other words, Ayres and Donohue are saying, be careful with using graphical analysis to try and draw causal conclusions becuase it is not always clear what is going on. This is why statisticians use multivariate analysis. A data series might be driven by 2 or more variables making it almost impossible to figure out what is driving the data. I think it is fair to say this is definitely the case with crime data. For example, in the quote above from page 11 is immediately followed by this statement

The main story is that robbery occurs more frequently in large, densely populated urban areas. Thus, one could not hope to establish the effect of a shall-issue law by looking only at which states have such laws and which don’t at any one point in time – a so-called “cross-section” analysis.

It is also possible that densely populated urban areas also recieve most of the resources for combating crime as that is where most of it takes place. As such if in the early 1990's there were policies enacted to combat crime in densely populated areas you'd expect to see more of a response in densely populated areas than in less populated areas. Could this be driving the various series? We don't know becuase Ayres and Donohue ignore this question so far...hopefully later they will address it.

But Ayres and Donohue press onward

But, there is no need to rely on visual inspection, since the statistical tool of regression can more formally and precisely do what the graphical analysis is trying to do – control for the initially different levels of crime and the common national forces acting on crime to see whether shall-issue adoption has any systematic effect on crime. We ran just such a regression model that controlled only for the average crime rate in the state and the common national influence each year and found that adoption of a shall-issue law was associated with an almost sixteen percent increase in robbery.28 Indeed, as Table 1, line 2 shows, running these same parsimonious regressions across all nine crime categories for the period 1977-99 yields results that, with only a single exception, are uniformly statistically significant and
positive, suggesting – however naively -- that shall-issue laws increase crime.29

Now I have a very serious problem. Table 1 does not have any coefficients listed in it save for violent crime (althought the standard errors are listed). I am also assuming the table I am looking at is table 1 in that it isn't labeled and it is the only one that could fit the description. Further, despite what the authors claim none of the standard errors are marked as being significant. It is sort of hard to judge the quality of research when the results of said research are not presented in the paper! I have to wonder what Tim Lamber would say if such a thing had happened in Lott and Mustard (1997). Still I suppose this isn't that bad since it the authors note that these regressions are simplistic and that there are a number of problems with them (omitted variables bias, the models do not appear to be robust, etc.).

Well that is enough for now.
Steve 10:35 AM

Wednesday, January 29, 2003

 
Instapundit's Popularity A while back several bloggers thought that Glenn Reynolds popularity was on the wane and that he was the victim of "market fragmentation". That is as more bloggers come online less people go to Reynold's site. The evidence, that Glenn's unique visitors had not reached its peak set in week 45 of 2002 at slightly over 240,000 vistors for that week. Well, I have bad news for these clueless "statisticians" Glenn just went slightly over 250,000 hits last week. Sorry guys, you should have paid more attention in your introductory statistics class.


Steve 1:27 PM

 
Iraq's and the Arab's Future This is an interesting article in Foreign Affairs about how the U.S. could use the war with Iraq as a chance to introduce modernism and democracy into the region. However, it is not a blindly optimistic article. For example the opening paragraph lays out what the reality will be in the event of a U.S. led war with Iraq.

There should be no illusions about the sort of Arab landscape that America is destined to find if, or when, it embarks on a war against the Iraqi regime. There would be no "hearts and minds" to be won in the Arab world, no public diplomacy that would convince the overwhelming majority of Arabs that this war would be a just war. An American expedition in the wake of thwarted UN inspections would be seen by the vast majority of Arabs as an imperial reach into their world, a favor to Israel, or a way for the United States to secure control over Iraq's oil. No hearing would be given to the great foreign power.

However, the author notes that the U.S. would be negligent if it did not try to ensure that modern and democratic institutions did not take root in Iraq.

The case for war must rest in part on the kind of vision the United States has for Iraq. The dread of "nation-building" must be cast aside. It is too late in the annals of nations for outright foreign rule. But there will have to be a sustained American presence if the new order is to hold and take root. Iraq is a society with substantial social capital and the region's second-largest reserves of oil. It has traditions of literacy, learning, and technical competence. It can draw on the skills of a vast diaspora of means and sophistication, waves of people who fled the country's turbulent politics and the heavy hand of its rulers. If Iraq's pain has been great in the modern era, so too, has been its betrayed promise. There were skills and hope that the polity could be made right, that the abundance of oil and water and the relative freedom from an overbearing religious tradition would pave the way toward modernity and development.

The author points to Japan after World War II and the U.S. occupation and transformation of the country.

If and when it comes, that task of repairing -- or detoxifying -- Iraq will be a major undertaking. The remarkable rehabilitation of Japan between its surrender in 1945 and the restoration of its sovereignty in 1952 offers a historical precedent. Indeed, the Japanese example has already turned up, in both American and Arab discussions, as a window onto the kind of work that awaits the Americans and the Iraqis once the dictatorship is overthrown. Granted, no analogy is perfect: Iraq, with its heterogeneity, differs from Japan. America, too, is a radically different society than it was in 1945 -- more diverse, more given to doubt, and lacking the sense of righteous mission that drove it through the war years and into the work in Japan.

Supposing that reform did take root in Iraq it could be just what is needed to put the region on a path towards stability, growth and peace. Add in the growing unrest in Iraq with its version of radical Islam and there are indeed some very good opportunities here. However, I do wonder if any in the current Administration have the presence of mind to see these opportunities and persuade President Bush to take advantage of them.

Update: Another thought occured to me. These opportunities and capitalizing on them could provide a nice long term strategy in the War on Terror. By establishing modern and democratic institutions in the Arab world would greatly undermine the influence of radical Islamists who preach "jihad" against the West and the U.S. in particular. It certainly would keep the radical Islamists from pointing the the situation of the people in the region and then blame it on a U.S. or Western conspriacy to keep them in that situation.
Steve 1:00 PM

 
The Indonesian Reaction to the State of the Union Apparently those Indonesians who watched the SOTU were not impressed and felt that the President did not make a convincing case for attacking Iraq. Is this important? I don't know, but Indonesia is the most populous muslim country in the world. Although according to the article nobody objected to an attack on Iraq because it too is a muslim country. I guess that is good news.
Steve 9:35 AM

Tuesday, January 28, 2003

 
Krugman Watch In his latest Op-Ed Krugman is claiming that Bush's popularity is slipping quickly.

But anyone who takes the trouble to look at the numbers knows that the thrill is gone. Mr. Bush's approval ratings have plunged over the last two months. A year ago he was, indeed, immensely popular; right now he's not significantly more popular than he was before Sept. 11.

Plunged? According to this poll he is down 12%, but most of that drop has been in the last few weeks not months. The Gallup Poll shows a long steady decline after 9/11. What the Hell is Krugman talking about? Beats me, but for somebody with his academic reputation to so badly misrepresent the empirical evidence is shameful.

Meanwhile, payrolls continue to decline; since the working-age population keeps rising, it's becoming ever harder for ordinary Americans to get jobs, or keep them.

Actually, according to the Bureau of Labor Statistics the unemployment rate is relatively unchanged when measured from December 2001 to December 2002. Not the greatest thing to point to, but the idea that payrolls are continuing to decline is not evident in the data. What is evident is that the economy seems to have stalled. Whether it will slide back into recession or pick up is open to speculation.

Overall, another poor showing by Professor Krugman.


Steve 10:23 AM

 
If It Is Only About the Oil How come we didn't suck Iraq dry of oil right after the Persian Gulf War? How come we are not going after a country such as Venesuela which is much closer (better logistically) under the pretense of the War on Drugs or something?

Further if we are going to limit ourselves to univariate explanations then what is the deal with France, Russia, and Germany? Is it just about the business they are doing with Iraq? After all, does the U.S. use 122 mm rockets? I don't think so...but don't the...Russians? Hmmmm.... Maybe these other countries like it the way it is just to sell weapons. Gee this monomaniacal focus on a single explanatory variable is fun and simplifies things soooo much.
Steve 9:27 AM

Monday, January 27, 2003

 
This Has to be a Material Breach Still I bet the Europeans back down.



Via the The Daily Rant

Steve 10:03 PM

 
Lambert's Weird Internet Behavior Many people have been going on about John Lott's use fo a psuedonym, Mary Rosh, for participating in online chats, noting it is rather weird. Well it seems that Tim Lambert has started a blog for Mary Rosh, but oddly claims Mary Rosh (a.k.a. John Lott) has started the blog. Its pretty obvious that it is Lambert's blog, but saying it is by Mary Rosh is a little weird...and slapping on the AOL e-mail account for Mary Rosh is also a bit unusual, IMO. I mean if your going to do something like make a website containing posts and comments by John Lott as Mary Rosh fine, but be more up front about it.
Steve 10:20 AM

 
Material Breach? The U.N. weapons inspectors will be filing an important report today and based on this article it doesn't look good, at least in my opinion.

Baghdad has complied only reluctantly with the UN inspectors

Iraq might still have stocks of anthrax

The country has failed to account for up to 300 rocket engines

Its weapons declaration last month contained no new material

The article also contains the following:

Any failure to co-operate has been identified as a possible trigger for war by Washington and London, but other key Council members say the inspections should be allowed to continue.

It would seem then that Iraq is in material breach and the sanctioning of military force would be justified. Of course, given the recent rhetoric coming from countries such as Germany and France, I doubt that the Security Council will sanction changing the toilet paper in the bathrooms. This next bit is from U.N. Resolution 1441:

False statements or omissions in the declarations submitted by Iraq... and failure to comply with and co-operate fully in the implementation of this resolution shall constitute a further material breach of Iraq's obligations

It seems to me that failure to hold Iraq accountable for its actions in this matter will cost the U.N. credibility (if it still has any). And people wonder why the U.N. is either ignored or laughed at. Generally the U.N. and its resolutions and declerations are meaningless unless supported by the military might of the U.S. I wont be the least bit surprised if the Security Council votes to give the weapons inspectors more time, which is a tactical victory for Saddam. Saddam is palying the waiting game hoping that as time goes by the desire for the use of force will wan...maybe he is hoping American Idol will capture the attention of the U.S. populace by and large sparring him from having to fully account for his weapons of mass destruction.

Update: Here is an article on how the EU Ministers are unanimous in their call for more time for the inspectors.

Here is an old article from Policy Review that has an interesting explanation as to the different policy responses by the U.S. and the Europeans. I think this metaphor summarizes the article very nicely

The psychology of weakness is easy enough to understand. A man armed only with a knife may decide that a bear prowling the forest is a tolerable danger, inasmuch as the alternative — hunting the bear armed only with a knife — is actually riskier than lying low and hoping the bear never attacks. The same man armed with a rifle, however, will likely make a different calculation of what constitutes a tolerable risk. Why should he risk being mauled to death if he doesn’t need to?

Europe is the man with the knife--weak and timid. The U.S. is the man with the rifle much stronger and less timid.

Here is another article that looks at the three primary "super systems" that are driving world events. The first is militant Islam, the second at at the opposite end of the spectrum is the United States with its individualistic liberalism (liberalism in the classical sense of the word), and with Europe falling in between the two in that if favors markets like the U.S. (to some extent), but also favors strong government (similar too militant Islam).
Steve 9:41 AM

Sunday, January 26, 2003

 
Lott Update--The Controversy That Just Wont Die Tim Lambert is keeping up with the daily updates. I find it amazing the guy has the time to do all that work! Anyhow it is the best "clearing house" for information on this issue.

A couple of points:

Tim writes:

Steve Verdon comes round and agrees that Lott should withdraw the 98% figure but thinks I should lay off now that I've won. But, Steve, Lott has not withdrawn the 98% figure.

Yes, he has not withdrawn the number and he may never withdraw it. Never the less, many people now consider that number highly doubtful and you have "won" in that regard. Tim also takes exception to Lott's claim of "adding a cautionary phrase in the second edition". Well to be honest both editions have conditional statements about the 98% figure. The first edition quote is

"If national surveys are correct, 98 percent of the time that people use guns defensively, they merely have to brandish a weapon to break off an attack."

And the second edition has

"If a national survey that I conducted is correct, 98 percent of the time that people use guns defensively, they merely have to brandish a weapon to break off an attack."

So these are not statements made with a tremendous amount of certainty. In fact, based on the what we have learned now, I myself would likely agree the number is not true and in future editions the number should be removed or at least amend that section to add that other surveys point to much lower percentages for brandishing a firearm as a crime deterrent.

Lott also has a response to the Donohue and Ayres paper, here. I'll take a look at it when I get done with the Donohue and Ayres paper.
Steve 9:28 PM

 
The Enemy Within Daniel Pipes makes the case for "racial" profiling of Muslims. He notes the following:

In the course of their assaults on Americans, Islamists - the supporters of militant Islam - have killed nearly 4,000 people since 1979. No other enemy has remotely the same record.

Islamists are plotting to kill many more Americans, as shown by the more than one-group-a-month arrests of them since 9/11.

While most Muslims are not Islamists and most Islamists are not terrorists, all Islamist terrorists are Muslims.

Islamist terrorists do not appear spontaneously, but emerge from a milieu of religious sanction, intellectual justification, financial support and organizational planning.--emphasis added

I emphasized the last one as I think it is an important point to always keep in mind. Most Muslims are not terrorists, and have no connection to terrorism. Aziz Poonawalla has made similar remarks when he notes that many of the practices that Western's object to are not derived from Islam, but from tribal customs that are not founded in Islam (although those who support and believe in these traditions try to appeal to Islam and its writings to justify their barbarism).

Anyhow a very interesting essay that I recommend.
Steve 9:03 PM

 
Game Theory and Equilibirium Concepts I read that essay by Mindless H. Dreck over at Asymmetrical Inormation and have a minor point of disagreement.

Dreck writes:

Are both sides pursuing a sub-optimal strategy? Nash would say no, this has been going on for years, they must be pursuing optimal strategies given the game in front of them.

Not quite. The strategies are sub-optimal in a global sense, but given the other player(s) strategy(ies) the strategy is optimal. This is different from optimality in a global sense in a subtle manner. Here is the difference, the payoff function for each player is denoted by U and the actions are x_i. Thus a globally optimal strategy is such that

U_i(x*_i,x*_j) > U_i(x_i,x_j) for all permisible x_i,x_j.

A Nash equilibrium on the other hand is such that

U_i(x*_i,x_j) > U_i(x_i,x_j) for all permisible x_i, and x_j is given (i.e. x_j is the other players Nash strategy).

Now, it follows that

U_i(x*_i,x*_j) > U_i(x*_i,x_j).

Hence it is usually the case that Nash Equilibiria are sub-optimal and that there is room for improvement, but that neither player can enforce the more optimal outcome. Overall this is a technical point of disagreement...really one of terminology. His analysis of the Palestinian problem is very interesting and the point Dreck makes below is I think a damn good one:

The cries from the Middle East that Palestine must be resolved before Iraq or other terrorist-sponsor states are truly "stasist" voices, attempting to preserve the status quo. They have put the cart before the horse, for no solution can hold without removing the pressure from other states who have an interest in perpetuating the conflict. That is why peace talks and accords have failed and might likely continue to fail. Talks and televised handshakes don't change the rules. It is the resolution or movement of other problems in the Middle East (no small order) that is most likely to cause the Palestinian issue to find a new equilibrium. Hopefully a peaceful one.

To use the language of game theory the peace talks and what not are "cheap talk" in that they are promises to adhere to strategies that have better outcomes for both players, but which have no enforcement mechanism. An example might help. Many of you might have heard of the Prisoner's Dilemma game. The game has two players and each player has a strategy of confess/not confess. The payoffs are in terms of years spent in jail (hence the negative number).

		Player Two
		Confess	Don't Confess
Player One	Confess	-5,-5	-6,0
	Don't Confess	0,-6	-1,-1

Now for a one shot game the strategy Confess/Confess is the (pure strategy) Nash Equilibrium. Clearly things would be much better if Player One and Two could agree not to Confess, but no matter how many handshakes, promises, threats, etc. each player will want to deviate from these "promises". That is all the talk prior to playing is not enforcable, i.e. it is cheap talk.

Now, when we aren't in a one shot game things get much, much more messy. If neither player knows when the game is going to terminate then this puts us into the realm of infinitely repeated games (even if the game has a finite end date, but that end date is uncertain we can still treate the game as one that repeated infinitely). When dealing with infinitely repeated games then there are possibilities for "punishing" bad behavior, and consequently one could arrive at a much better payoff arrangement. The problem here is which equilibiria do you select? One of the problems is that there is a plethora of equilibria. This follows from an old result known as the Folk Theorem (it is called the Folk Theorem because it has been around a long time and is part of the "folklore" of game theory). Anyhow the Folk Theorem says, in short, that if an outcome is better than the single shot equilibrium then it is pretty much an equilibirium in a repeated setting as well. Now before you look at our example here and conclude that it means people will thus play Don't Confess/Don't Confess it also means people might periodically deviate as well to learn what sort of strategy their opponent is using to see if they can improve their own situation by deviating periodically. So there are in reality a very, very large number of equilibiria and in some cases infinitely many. This is why the game theory literature has so much on what are called "refinements" which is a way of reducing the number of permisible equilibiria.

What does all that mean with regards to the Palestinian/Israel situation? Well, I think it is safe to say we are in a repeated game, but we also have more than two players. Anybody who thinks this is just a problem between the Palestinians and the Israelies is very much out of touch. Also, on the Isreal side of the game it isn't necessarily the same player. I know it is often fashionable to think of nations as just one big person, but I have bad news they aren't! President Clinton is not in the White House anymore and his approach to the problem was different than President Bush's approach. Further, given Dreck's excellent point that ignoring the other players destines the peace process to failure, perhaps President Clinton's approach was the wrong one. Maybe, a secondary benefit to ousting Saddam Hussein will be the removal of one more aggitator in the Palestinian/Isreali conflict
Steve 8:05 PM

Friday, January 24, 2003

 
Tim Lambert Changes his Mind Today Tim writes:

The most important thing about this affair has always been whether Lott's 98% brandishing claim is true.

I find this hard to believe. If this were simply the case, all you'd need to show that the number is not that credible is the sample size. Once you have that you're pretty much done. Since Lambert has had this information since January 16th I am not sure what the fascination is. Lott's current survey while consistent with his earlier number is also consistent with results that have a much higher percentage where the gun owner had to fire the gun while using it defensively. Frankly, this is enough for me to discount the 98% as being true. Couple that with other studies that come in with much higher percentages and I'm pretty convinced. The data necessary for me to change my mind now will have to be alot more impressive. So, if this was the goal...why continue with this? You're done, pat yourself on the back. Obsessing about it is not going to make you look good.

On another issue: Note to Tim...just reading the abstract of an article is not sufficient, IMO. I have had plenty of occasions where an abstract looked plenty promising only to find out the article was of little help. Of course, you can only make this assessment after reading chunks of the article at the very least. So I'll stick with the reading of Donohue and Ayres article vs. just going with its abstract.

Kevin Drum is starting to go beyond this to conclusion that all of Lott's work is junk. Careful there Kevin you are just short of also impugning the character and research of David Mustard. After all, the Lott & Mustard paper has David Mustard as a co-author. While the 98% figure is very dubious and should be retracted, IMO, concluding that all of Lott's research is the result of sloppy work and or research has much broader ramifications and one should exercise some restraint.
Steve 4:12 PM

Thursday, January 23, 2003

 
Blogging My blogging might be a bit sporadic at times. I have finally gotten Moveable Type uploaded to my own domain and I am working on getting my new blog template ready so I can cut free of blogspot. So, if I post lightly or not at all on some days, sorry...I'm working on something that will hopefully result in less frustration on your part and mine...i.e. not having to contend with blogger/blogspots sensitive servers and haloscan's all too frequent hiccups.
Steve 11:41 PM

 
Tim Lambert Okay, I admit I have some problems with some of his statements, but I have to give the guy credit for doing a simply amazing job of collecting all sorts of links on the Lott Survey fiasco. He is really on top of it in terms of finding sites and commentary form all over the place and from different sides. If this subject interests you and you don't know about Tim's webpage on this issue (how this could be I don't know) then check him out and book mark it for future reference.

Regarding Lott and his survey, I must confess some disappointment (okay alot of disappointment) in that it sure does look like the sample sizes for the survey are too small for the results to be of any serious value. Granted it is tangential to his main thesis, but it does provide a chink in the armor so to speak. People can now try to fit the thin edge of a wedge into that chink and go from there to claiming his research on crime rates and shall issue CCW laws are discredited. This is not the case (i.e. the logic is bad, although it could still be the case that Lott's main claim that simply put shall issue CCW laws reduce crime is wrong).

Anyway, I plan on reading through some more of the Donohue and Ayres paper and posting on it tomorrow if I can.
Steve 11:34 PM

 
The Race Wars When discussing race and genetics one of the things that will periodically crop up is the following claim:

The variation within a "race" is greater than the variation between races, genetically speaking; thus races do not exist.

Personally as somebody with some small amount of training in statistics and mathematics I find the above specious. Sure it sounds great, say it at a rally and you'll likely get deafening applause. But lets strip away the racial aspects. Suppose we have two random variables X₁ and X₂. Now, we can calculate a variance-covariance matrix for these two variables

s₁ s_1,2
s_1,2 s₂

Now that quote

The variation within a "race" is greater than the variation between races, genetically speaking; thus races do not exist.

is saying, to me, is the following

s_i,j > s_i,

and that based on this we cannot make any meaningful distinction between X₁ and X₂. What is the problem? To be able to calculate the s_i,j and s_i you have to be able to distinguish between X₁ and X₂ which supposedly we can't do. In other words, the claim is internally inconsistent and...well stupid. Further, I don't know of why this condition

s_i,j > s_i,

means you can't suddenly distinguish between X₁ and X₂.

Why do people make this claim then? I don't really know. My guess is because it is not a popular position to take. Imagine if you were a professor of biology at a university and you made the following claim:

There are measurable genetic differences between the races.

How long till there are angry Op-Ed columns? How long till Josh Marshall and Atrios are writing nasty little blog posts about you and digging into your past looking for something truly vulgar? How long till you have to go out and publicly flagellate yourself till those who find the statement offensive are appeased? What about your funding and future prospects for publication?

What makes this truly amazing is you see the genetic differnces. When you see a person of the black "race" you see a darker skin pigmentation which is determined by genes. When you see a person who is asian you see somebody with different texture hair and differently shaped eyes, which is controlled by genes. Another fear of those who abhore the above view is that perhaps they fear that not only do these genetic differences imply phsyical differences but possibly behavioral ones or differences in intelligence. Maybe they do, I really don't know...nor do I give a shit really. Just because you admit that there might be genetic differences between races doesn't mean you have to suddenly stop judging individuals as individuals. Also, you are not suddenly duty bound to make decisions based on an average of each race. If you do feel duty bound to follow such a policy then I also expect you to gamble everything on a single throw of the die at the craps table in Vegas (well if you are logical and consistent that is). So we are different, we are different as individuals as well as races, personally I find it a good thing. If we were all the same it would be a very dull place and a world without Pakistani food and Vietnamese food is one I don't want to live in.

Update: An e-mail to me about this post pointed out that some examples of those using this argument would be a good idea to avoid the obvious response of, "Steve that is a strawman argument." Okay, good point. Here is one such example. And here is another. And yet another site utilizing this argument. The response to this, IMO is this:



Of course the scientist who generated that map, Cavalli-Sforza, has had to use some pretty dumb rhetoric, IMO.

It is important, however, to note that current classifications depend on external appearance, which is due to very few genes (hereditary factors) affecting skin, hair and eye color. Body and face size and shape may involve a few more genes but, like the former, are the result of an adaptation to climate (including diet and customs, which obviously also depend to a large extent on climate). Common belief in the "existence" of races must depend on local uniformity of skin color in different environments: dark in tropical climate, brown at some distance from the equator, light brown in south Europe, reaching the highest degree of whiteness in the southern Baltic.

I think Mr. Cavalli-Sforza is attacking a simplified position, i.e. that races are genetically "uniform". We already know that individuals are not uniform so it is pretty much a given that races have to have uniform skin color. If we change the definition to one of distributions this counter argument fails. But this also brings up another argument against the existence of races: whatever differences there are it is due to only a "small number of genes". Fine, but even a single gene difference can be significant. Sickle Cell Anemia is due to one gene. Further, that single gene difference is also highly correlated to skin color. Basically, there is alot of work going into trying to prove that races don't exist....how you can find evidence for something that doesn't exist is beyond me.
Steve 10:57 PM

 
According to Ed Leamer the U.S. Economy to Remain Sluggish Ed Leamer the director of UCLA's Anderson School's forecast predicts that the U.S. economy will remain sluggish for the next 6 months. Also, that there might be an increase in business investment it will be countered by a drop in consumer spending brought on by consumer debt loads.

Like the unique “business recession,” we are currently experiencing a unique “business recovery,” as the businesses try to sell off inflated inventory and in particular get more life out of technology investment made in the late 1990s. When business begins to increase production (due to reduced inventory) and start purchasing technology again, then the economy will pick up steam.

Steve 10:08 AM

 
Predator or Prey This is an article about the Predator Drones that have become one of the preferred weapons in the war on terror. This article, as might be guessed from the title, is critical of the these new weapons. The drones are extremely easy to shoot down, according to the article, and that over half of the U.S.'s fleet of 60 or so Predators have been shot down (although on the upside its only drones that have been shot down, i.e. no Americans were at risk). Perhaps Shachtman's comparison to the Patriot missile from the first Gulf War is apt.
Steve 10:02 AM

Tuesday, January 21, 2003

 
CalPundit's Myopia Kevin Drum is confused as to why some "conservative" bloggers are upset that liberals went to the A.N.S.W.E.R. rally. He notes that sure, some of the top people at A.N.S.W.E.R. are part of the Workers World Party which says nice things about North Korea and Cuba, but he still doesn't get it why some people might be upset.

How about this Kevin, you and other liberals did the "guilt by association" trick with Lott and racism and Republicans. Perhaps now people are just returning the favor. Some of Kevin strongly implies that Republicans are racist (or at the very least appear racist). Here, here, and here. The basic message, if your a conservative then racists like you and therefore you are a racist too.

Others have jumped on this band wagon too. J. Bradford DeLong jumped in with both feet.

Maybe next time cooling it on the guilt by association thing would be prudent.
Steve 10:41 PM

 
Dubious Data Awards for 2002 There are some pretty funny ones in there. At the bottom Dennis Byrne writes:

Not to be outdone, I've created my own award, the U.S. Rep. Charlie Rangel trophy, which I present to myself and other pundits who swallowed the conventional wisdom that disproportionate numbers of blacks are at risk in the military or were killed in the Vietnam War.

A Defense Department study found black fatalities in Vietnam were about 12 percent of all Americans killed--proportionate to the number of blacks in the general population then, and much lower than the proportion of blacks in the Army.

Link via Henry Hanks at Croooow! Blog.
Steve 2:53 PM

 
Ooops, Lott does Have Some Evidence Well, looks like Lott is starting to come up with some proof that he did do the survey. I wonder if people like Kevin Drum are going to comment? I doubt it.

Update: Ooops, I spoke too soon and have to eat some crow. Kevin Drum at CalPundit does comment although it isn't that glowing a comment and would rather a student who helped conduct the study would come forward rather than a respondent.
Steve 11:19 AM

 
Krugman Watch Krugman once again launches into a tirade against the evil Overlord, a.k.a. George W. Bush (note: send President Bush this list). Anyhow, Krugman bashes Bush for being dishonest and stupid.

For example: On Saturday, in his weekly radio address, George W. Bush declared that "the tax relief I propose will give 23 million small-business owners an average tax cut of $2,042 this year." That remark is intended to give the impression that the typical small-business owner will get $2,000. But as the Center on Budget and Policy Priorities points out, most small businesses will get a tax break of less than $500; about 5 million of those 23 million small businesses will get no break at all. The average is more than $2,000 only because a small number of very wealthy businessmen will get huge tax cuts.

Well...as they say, there are liars, damn liars and Paul Krugman. We still don't know anything about which companies are getting the tax cuts. Are those who are not getting any tax cut small time "mom & pop" operations? Are they owners of larger companies barely making the "small business" cut off? If it is a small number of very wealthy businessmen getting a huge tax cut...couldn't we get at least one example? Guess not. So Krugman really hasn't given us much more information than President Bush has given us. Yes, the average is sensitive to large observations in the tails and the better measure is the median...but didn't the Center on Budget and Policy Priorities calculate a median?

Oh great, even the Center for Budget and Policy Priorities gives very little information on this. Looks like they are referencing a study by the Urban Institute...with no link (what is it with these nimrods, can't they code in a freaking link?). Also, one has to remember that 50% of the population pay only 4% (approximately) of the federal income tax. Hence, there are going to be lots of people you can't cut federal income taxes for because they don't pay federal income taxes. The Krugman types don't tell you that, they just tell you that x% are getting no tax cut at all. Well, if you are not paying any federal income taxes...how much of a tax cut can you get? Oh...wait, you mean you want a federal income subsidy. At least be honest (oops, I forgot this is Paul Krugman we are talking about).

And bear in mind that the budget deficits of state and local governments are forcing cuts in medical care for the poor and public services for everyone. Many states, even those with Republican governors, will be forced to raise taxes too - but the burden of those increases will fall on the middle class, not the rich.

Yes Prof. Krugman this is correct...but howcome you are not railing against the evils of the highly regressive sales tax, excise taxes, and so forth? Hmmm?
Steve 11:14 AM

 
Shooting Down the More Guns, Less Crime Hypothesis--The Theory Ayres and Donohue go over the theory of the "more guns, less crim" hypothesis in section I of their paper. I must confess to a tremendous amount of disappointment here in that the "theory" is only about 2 pages long and consists of little more than some speculation on why the Lott and Mustard hypothesis may be wrong. Ayres and Donohue layout the theory of the Lott and Mustard hypothesis in one paragraph. The idea is fairly simple, that as the number of concealed firearms carried by law abiding citizens increases the risk of perpetrating a crime on these citizens rises. This acts as an effective deterent to criminal activity. Further, with concealed weapons there is a spillover effect. The spilloever (or positive externality) arises from the fact that a criminal cannot tell who is carrying a gun. If the criminal had some sort of signal (or could tell) who was carrying a firearm the criminal could simply avoid such individuals and focus on those not carrying firearms. Thus, carrying concealed firearms provides a general deterent to criminal behavior.

The authors offer several possible replies as to why the above effect might not hold, or more accurately might not dominate. The first is that the laws that Lott and Mustard were analyzing have very simple criteria for obtaining a carry concealed weapon (CCW) permit. As such many "angry individuals" might get such permits who most of us would prefer did not have such permits or even access to a firearm. Aryes and Donohue note that Lott and Mustard's reply that the number of crimes committed by CCW permit holders is very small. Aryes and Donohue speculate this could be much higher though as not all criminals are identified and linked to a specific crime. I can't help, but note that Tim Lambert, who sites this article heavily, also makes this same mistake. That is Lambert relies solely on the number of reported incidents and ignores those incidents that were not reported.

There were only 12 incidents where a criminal encountered an armed permit holder. Compare this with the roughly 100,000 violent crimes in Dade county in that period. Clearly the chance of a violent criminal encountering an armed victim increased by at most 0.012 percentage points.

This is incorrect if the number of incidents involving a criminal and a CCW permit holder is larger than 12 and only 12 are reported. Lambert provides us with the above even when he knows it is likely to be too low of an estimate:

The true figure is considerably less since some permit holders may have carried legally or illegally before the law, only half of the 12 incidents involved defence against a violent crime, and crimes where are a gun is used defensively are more likely to be reported to the police than crimes in general. (NCVS data [38] indicates that 65% of crimes where a gun is used defensively are reported to the police, compared with 43% of crimes in general. Kleck's survey [22] indicates that 64% of defensive gun uses are reported to the police.)

In other words, using the 65% figure the actual number of incidents involving a CCW permit holder would have been about 15 (maybe more or less depending on the ranges of the 65% estimate). Clearly Lambert is guilty of sloppy statistics.

Another possible response to the Lott-Mustard hypothesis is that it could result in an "arms race" between criminals and permit holders. Further, such a law might also induce a criminal to shoot first, then look for the wallet or valuables. Thus, the CCW law could increase the number of violent crimes.

The third response is that the injection of a gun into a minor dispute could turn the dispute into a lethal one. From this the authors slide on into the notion that again many of the people who have guns are less than model citizens and cite one example of a known criminal getting a permit. Of course, the authors ignore the possibility that the person in their example might very well have carried illegally as well. It seems that the rigorous logical standards apply only to Lott and Mustard and not to themselves.

The authors also mention the theft of guns and that more guns in households (due to the shall issue CCW law) would induce more burglaries of residences as a stolen gun can be sold quickly and for a sizeable sum. I am struck at this point by Tim Lambert's silence on this part. He has criticized Lott and Mustard by claiming that there was not an increase in the number of guns in the states that passed shall issue CCWs. If this is true then Lambert, if he were not so biased, would note this criticism applies to Aryes and Donohue as well as Lott and Mustard. But, when you are on a mission who cares about bias and logical consistency.

The last point is almost silly. The authors note that having a shall issue CCW law means more work for police officers. The police officers will have to now check that the firearm is indeed legal and that the person does indeed have a CCW permit. I find this complaint ridiculous. It holds for other things as well that can and are used illegally, automobiles. Last time I was stopped by a police officer they went to the trouble to verify that I was who I said I was, that the registration was current and that the car was registered to my wife, and that the insurance was up to date. Should we get rid of cars for this reason? Further, it is the job of police officers to make sure that laws are being followed and that when they are not being followed they either issue a citation or arrest the offender. The authors also state

As it stands now in Illinois, anyone caught with a gun in public is violating state law and can be immediately brought into custody without the further need for investigation, which the state police believe has been a powerful tool for taking criminals off the streets.

Seems this state law also make work for the police. Further, I wonder how many otherwise law abiding citizens were arrested under such a law and what the overall social impact is.

In short, I found this section horrid in terms of a "theory" section. Basically I see it as Aryes and Donohue's musing jotted down on paper. For example just prior to this section Aryes and Donohue were complaining that there was no theoretical justification for the rise in property crimes. Yet they themselves provided a possible theoretical explanation themselves! Further, the idea of subsititution in crimes seems to have completely eluded them. The idea here is that given an increase in risk with crimes on people the criminal migh switch to property crimes that carry less risk. Overall a very, very weak section of the paper.
Steve 10:20 AM

Saturday, January 18, 2003

 
More on the Case Against Intellectual Property After the Supreme Court Decision in Eldred v. Reno I posted this link to the work of Michele Boldrin and David Levine on how intellectual property right laws might actually be impeding growth not helping it.

This position flies in the face of conventional wisdom on intellectual property rights. The conventional wisdom is that ideas are non-rivalrous, which means that one person’s use of the non-rivalrous good does not diminish another person’s ability to use the good. Consider the idea embodied in the Pythagorean theorem, even though I know it does not prevent you from learning it and using it. Now the problem is that those who come up with good ideas probably wont be compensated for their idea and hence may not share it, or even worse wont even devote any resources to coming up with the idea at all. So, to get around this problem, society decides to give the creator of the idea a temporary monopoly on the idea. Even though this is not the most efficient outcome it is the best we can do (in economics we call this outcome the “Second” Best).

Boldrin and Levine argue that today intellectual property rights laws have gone well beyond the idea above. In their words

Economic Efficiency, and common sense, argue that ideas should be protected and available for sale, just like any other commodity. But “intellectual property” has come to mean not only the right to own and sell ideas, but also the right to regulate their use. This creates a socially inefficient monopoly, and what might be better called “intellectual monopoly.”

The authors use a simple example to explain their point. You can buy a potato and when you buy it you can eat it, throw it away, plant it, or make it into a sculpture. With say computer software and current laws you do not have these freedoms. Even though you have bought and paid for the program, i.e. it is now your property; you cannot do many of the things you would do with other commodities you have bought.

Another interesting point raised by Boldrin and Levine is the following. Economists generally frown on monopoly. Monopolies increase prices, decrease output, impose a deadweight loss, and some even argue reduce innovation. But, to promote innovation and creativity we must give the innovators and creative monopoly power. Seems like a paradox. However, the conventional argument is that the cost of innovation is a fixed cost and the marginal cost is constant (perhaps zero). Given that perfect competition results in prices equal to marginal cost, the fixed costs are not recovered and hence competition means those who innovate wont recover their fixed costs. So to get around that, give the innovator a temporary monopoly.

Boldrin and Levine argue that the cost of creation is a sunk cost. What is the difference between a sunk cost and a fixed cost? Fixed costs are costs that do not vary with the amount of output. Suppose you want to produce a commodity and you rent out a building to be your factory. Now if you produce 1 unit of the commodity or 1,000 units the rent is still the same. Hence it is a fixed cost. Sunk costs on the other hand are those costs that once committed to are unrecoverable. Suppose you are going to a resort and you are halfway there. You have paid $100 in non-refundable deposit. You also realize you don’t want to go to the resort that weekend and would be better off staying at home. Some people say, “Go to the resort or you’ll have wasted your $100.” The problem with this is the $100 is gone no matter what you do (keep going or turn back). It is a sunk cost. Another way to think of a sunk cost is any cost you must pay regardless of whether or not you produce anything or not.

If the costs associated with innovation are sunk costs then they should not be part of the calculus in terms of pricing, as they are unrecoverable no matter what you do. Further, sunk costs are no unique to innovation and intellectual property. Growing potatoes entails sunk costs, yet Congress is not passing laws to give the potato growers the rights to downstream licensing (you can steam and fry your potatoes, but just don’t bake them or plant them).

Another problem with the current laws for intellectual property is that they are expensive. They are expensive to enforce. Look at all the effort the recording industry is going to try and kill the file sharing programs. Also look at all the effort that goes into trying to prevent people from accessing the information on CDs and DVDs. Further, some of the newer ideas to try and enforce the downstream licensing agreements are extremely invasive in terms of privacy.

Then there is also the rent-seeking behavior which is also costly. Rent seeking is especially wasteful in that it does not entail new production, but alterations in the current economic “pie”. Eldred is a particularly fine example of rent seeking in that it applied retroactively to intellectually property that already exists. Since these “innovations” already exists further extensions on their property rights clearly does not result in their productions…since they already have been produced. This part of the case was purely about redistribution of economic resources. Further, it flies right in the face of the “conventional wisdom” of intellectual property rights. Remember the argument goes thusly. We all want to enjoy the fruits of innovation, but without rewarding the innovator we wont have the innovation. So we will give the innovator a temporary monopoly, and when that monopoly is over we can all enjoy the fruits of the innovation cheaply. But along comes Congress and significantly alters this game plan.

So these are some of the problems with the conventional approach to intellectual property. But Boldrin and Levine don’t stop there. They put forward a very simple model that looks at the first sale price of intellectual property when there is no downstream licensing. The conclusion is that when demand is elastic that when there is technology that allows for faster reproduction, the first sale price goes up. Hence, something like Napster would actually make the first sale price higher for the creator of intellectual property in a competitive market. That last part is critical. Currently the market is not competitive, but a monopoly. If you like the Red Hot Chili Peppers there isn’t much you can do if you think the price of their latest CD is too high. Further, since it is a monopoly the profits will be higher than in the competitive case, hence the opposition to file sharing by the recording industry.

Anyhow, that is the basic argument that Boldrin and Levine have put forward against the current laws for protecting intellectual property.

Steve 7:07 PM

Friday, January 17, 2003

 
Bruce Sterling Video This guy is in full rant mode and boy is he hilarious.
Steve 11:32 AM

 
Click the Following Link Only if You are Patient or Have Broadband It is a four minute film by the Marine Corp on Afghanistan. Via Instapundit.
Steve 9:55 AM

 
Preferential Treatment in University Admissions I see there is lots of people going on about the Bush Administrations opposition to the admissions policy at the University of Michigan. As I understand the policy, each student is scored based on their application package. A perfect S.A.T. garners 12 points, but having the right skin color gets you 20 points. Now the right skin color is black or hispanic...but not Asian. Why not? Asian suffer from racism. They have been treated like shit by this country in the past. Now the usual suspects are saying that the Republicans and Bush are opposed to Civil Rights. Of course this is a load of bravo sierra. Ask yourself this question.

Would this system be racist if whites and asians got 20 points for their skin color and blacks and hispanics did not?

If the answer is yes, then the policy is racist no matter if you switch whites and asians with blacks and hispanics.

Now some of the supporters of these types of systems say all they want is a level playing field. I have bad news for these people, the playing field is never level. Further, is it a question of equal opportunity or equal outcomes? If it is equal outcomes then howcome you are not supporting a redistribution of all wealth in the U.S. Total it up, divide it by the population and then give each person that much money. Well, why don't you support that equal outcome?

One of the reasons the playing field is not level is because the Asian students bust their asses by and large. I think it is fair to say that culturally Asians tend to put a much higher value on education than other racial and ethnic groups. I know some of you might be thinking, "RACISM!" But lets consider the fact that the number of Asians at univeristies such as U.C.L.A. and U.C. Berkeley is way, way, way above their proportion of the population.

Basically the idea is the University of Michigan doesn't want "too many" Asians and Whites to be at their school. That is what their policy says. It says, "Okay, you beat this guy in terms of academics and other "objective" requirements, but because you have the wrong racial or ethinic background your not going to get in."

Other opponents argue that this type of policy is okay because of legacy admissions. Fine, you are absolutely right stop that crap too. There, are there any other real arguments that can support this type of admissions policy?
Steve 9:32 AM

Wednesday, January 15, 2003

 
Instapundit's Popularity Revisited Or why some people are clueless about statistics. Remember that big flap about how Instapundit's popularity was sliding? That he hadn't hit his 20 week peak for something like 8 weeks? Well, Instapundit still hasn't hit that peak...but it sure was close last week! 225,265 hits in the second week of January vs. 240,611 for week 45 of 2002. Lets also ignore the data from weeks 49, 50 and 51. Only by ignoring all those data points (leaving only 6 data points) do you get a downward trend. Yes, chuck 40% of your observations to get the conlcusion you want. Lies, damn lies, and statistics.
Steve 2:49 PM

 
Was Napster is Right? The Supreme Court has upheld the Bono Act (via Instapundit) which extends the amount of time on copyrights. As such I thought it'd be good to post that link that gives an overview of the work by David Levine and Michele Boldrin on intellectual property rights. Levine and Boldrin are currently putting forward the case that intellectual property rights actually stifle innovation and economic growth. Needless to say this flies right in the face of conventional wisdom on this issue.

Update: Robert Musil has a post on this that is well worth a read. Musil notes that the economic arguments filed by those opposed to the extension were exceedingly simple, basically some discounted value calculations. He also notes, that setting the time limits on copyrights is a power delegated to the Congress by the Constitution so there wasn't much the Supreme Court could do. Levine and Boldrin have a more sophisticated argument, plus some data to back it up.

Link

That is a graph of per capita literary works. Clearly there hasn't been an increase in the production of literary works per capita even though the time a literary work is copyrighted has increased dramatically.

Levine also notes the absurdity of the production claim with regards to copyrighted works that already exist. Since they already exist the claim that extending the copyright period will encourage their production is false since they already exist.
Steve 1:54 PM

 
Shooting Down the More Guns, Less Crime Hypothesis This paper is Tim Lambert's pet study in his crusade to debunk John Lott's research. As such I decided to give it a much closer look. It is pretty darned big though, 55 pages with lots of statistical results and such so I'll periodically be posting bits and pieces here (latter I'll try to summarize these disparate comments). My first observation comes from the introduction of the paper:

We conclude that Lott and Mustard have made an important scholarly contribution in establishing that these laws have not lead to the massive bloodbath of death and injury that some of their opponents feared. On the other hand, we find that the statistical evidence that these laws have reduced crime is limited, sporadic, and extraordinarily fragile. Minor changes of specifications can generate wide shifts in the estimated effects of these laws, and some of the most persistent findings -- such as the association of shall-issue laws with increases in (or no effect on) robbery and with substantial increases in various types of property crime -- are not consistent with any
plausible theory of deterrence.

Not exactly something you'd think an anti-gun advocate like Lambert would point to for support. The problem is that even if Lott and Mustard's work is not robust (i.e. minor changes in specification cause noticable changes in the model parameters) and is badly specified it does not then follow that the opposite conclusion is true. In fact, the opposite conclusion, i.e. a blood bath or an increase in crime, does not hold according to what I have read of Ayres and Donohue so far. At best you can say that mor analysis is needed.
Steve 9:58 AM

Tuesday, January 14, 2003

 
Well I'll be Dipped in Shit A thread at Democratic Underground where most of the posters (I have read so far) haven't had the typical knee-jerk reaction:

CAPITALISM IS BAD, BAD, BAD!!!

Of course, there is some crap there.

This comment is one

In fact, most of Marx's early writings were based on the works of Adam Smith. Marx was more of a social commentator than an economist anyway -- he specialized in pointing out the PROBLEMS of the industrial economy and proposed some broad solutions.

Actually, Marx' economics was more in line with that of David Ricardo's. Granted Ricardo didn't repudiate Smith wholesale, but his work was original and not just a copy of Smith's work.

IrateCitizen's notion of corporatism is just weird. If the U.S. were corporatist (which is a modern form of mercantilism...why it is modern I don't know IrateCitizen doesn't explain other than nation states doing things to benefit corporations) why pursue policies of free trade? Free trade increases competition. Perhaps if he had pointed to the U.S. recent steel tariffs and the protections on agricultural products he'd have a better case.
Steve 1:23 PM

 
Molly Ivins Molly, needless to say, is very, very unhappy with the new Bush tax plan. First she notes this:

-- According to the Urban-Brookings Tax Policy Center, the effect of eliminating dividend taxation is that the average benefit for those making less than $10,000 would be $6, and average benefit for those making more than $1 million would be $45,098. Quick, high-schoolers, let's practice up for the those SATs by figuring out by what percentage $45,098 is bigger than $6.

Lets also not forget that the person earning $10,000 will only pay $231 in taxes. Assuming this is somebody who gets paychecks every two weeks this means the federal income tax withheld from his paycheck is $8.89. Gee, how much of a tax break can we give this guy? Suppose it is $231 is it going to be much of a stimulus? No. This kind of income is what you'd see with minimum wage workers. How many minimum wage workers are there? Let check the Bureau of Labor Statistics. In 1990 the number of minimum wage earners was estimated to be 3.9 million. Now lets ignore the fact that many of these are part time jobs and assume that all 3.9 million are full time "career" minimum wage earners and Bush eliminated their entire tax burden. How much would this put back into the economy? How does $900,900,000. For an economy well over $6 trillion in GDP that is not even a drop in the bucket.

Yes I know, when you factor in state and local taxes it isn't quite so disparate. The problem here though is that state and local taxes are usually collected via highly regresive means such as sales taxes, excise taxes, and property taxes. My response is so what? This is not something Bush can address and even getting rid of the federal income tax for these low wage earners would not do much to reduce the regressiveness of the system. The $231 the poor wage earner pays above is only 2.31% of his income. Suppose after all taxes are factored in the guy has $9,000 in net income (after taking out Social Security, Medicare, and other taxes). Now he is probably going to have to spend almost all of that, thus it will be subject to the state sales tax. Since not all expenditures are subject to sales tax, lets say he pays $361.25 in sales tax. This is not something Bush can address, it is up to each state's governor and legislature to address.

Molly continues to point to the "unfairness" of Bush's proposal with this

-- Bush also wants to accelerate the income-tax cuts slated for 2006. Look at this folly. The top 5 percent of taxpayers would get 70 percent of the benefits on that one. The bottom 80 percent would get 6.5 percent of the benefits. Ditto with accelerating the 2004 tax cuts: 64.4 percent to the top 5 percent of taxpayers; 7.7 percent to the bottom 80 percent.

Time to get out the chart again...

Income Percentile	Taxable Income	% of Total US Income	% of Total Income Tax Paid
Top 1%	$313,469	20.81%	37.42%
Top 5%	$128,336	35.30%	56.47%
Top 10%	$92,144	46.01%	67.33%
Top 25%	$55,225	67.15%	84.01%
Top 50%	$27,682	87.01%	96.09%

From this chart we can see that the bottom 75% pay just under 16% of the federal income tax. There just isn't much left to cut. I suppose we could cut it all, but it still wouldn't amount to much. I guess its been sometime since Molly has been to elementary school. Actually, what Molly really wants, but doesn't have the intestinal fortitude to come out and state explicitly is income redistribution. She just doesn't want a tax cut for the poor, but wealth transfers. Take the money from those at the top and give it to those at the bottom.

Speaking of damn lies and statistics, one of the little games being played in Washington is that the Republicans want to switch to Enron accounting on the economy. They're leaning on both the Congressional Budget Office (CBO) and the Joint Committee on Taxation to change the way they make their economic estimates. According to the R's, "static scoring" -- as opposed to your "dynamic scoring" -- overestimates the cost of tax cuts by ignoring their role in boosting economic growth. Why, claim the R's, tax cuts pay for themsleves! If that's so, why are all the states going broke? Bring on Arthur Andersen and mark-to-market accounting -- that'll perk up the economy.

Gee Molly, maybe it has to do with bad management. I am not familiar with the situations in other states, but here in California the problem can be laid squarely on the shoulders of the politicians. There haven't been any tax cuts, but the idiot politicians decided to go on a spending spree with the dotcom tax revenue like drunken sailors on shore leave looking for hookers. The result: a massive budget deficit of about $35 billion over the next 18 months. Oh and lets not forget the orgy of spending on expensive energy contracts Davis entered into at the height of the energy crisis. Lets not mention that those energy contracts are now practically worthless. Nope, lets just blame it all on tax cuts. Also, lets point out that not all the states are going broke. I can just imagine Molly writing that column with her breathing getting faster and faster and banging harder and harder on the keys. Apparently all that oxygen resulted in her becoming dishonest.

The only good part of the Bush's tax cut plan is the $400 increase in the tax credit per child -- at least that spreads it around a little. Naturally, that's the one part of the plan right-wingers hate.

I think the only person who hates is you Molly. I think you hate Republicans. Your inability to tell the truth at all on this subject indicates a pathalogical hatred. I don't know why, maybe a Republican ran over one of your cats or something.
Steve 10:22 AM

 
Krugman Watch Krugman gets an 'F' with this latest column. He writes

My colleagues on the editorial page dubbed the Bush administration's proposal to eliminate taxes on corporate dividends "The Charles Schwab Tax Cut." Indeed, the idea seems to have originated in remarks Mr. Schwab made last summer in Waco. But a closer look suggests that it should actually be called the "Tax Complication Act of 2003": it will do little if anything to create jobs in the economy as a whole, but will be a bonanza for tax lawyers and accountants.

The part about jobs I am not so sure about. Sounds more like an empirical question and hence is another example of Krugman talking out of his ass. The latter part I agree with (hence 50%...or an 'F'). Complications in the tax code are not good. The problem is that these complications make it more expensive for people to comply with the tax code and given that a complicated tax code is no defense against an incorrectly filed tax return people really are motivated to pay such costs.

His closing comments reminds me of what Daniel Drezner said about Krugman

Now, this is certainly not strident. It does border on megalomaniacal paranoia, however.

I think some of Krugman's colleagues need to go in there and stuff some more aluminum foil into his hat.
Steve 9:26 AM

Sunday, January 12, 2003

 
CalPundit and Statistics While cruising the Daily Rant I came across that link. Kevin is commenting on what appears to be some very questionable results by John Lott in his book, More Guns, Less Crime. Apparently Prof. Lott has mistated some results of a survey and the corroborating is not available. What gets me is Kevin's closing comment:

UPDATE: I long ago decided that there was probably not one single evenhanded study in the entire literature on gun control, so I decided to simply stay agnostic on the whole empirical question of whether guns cause crime, prevent crime, or have no relation at all to crime. However, for those interested in a rebuttal to Lott's work, you can find one here.

In other words, because there is alot of bad statisitcal analysis on guns and gun control don't bother looking for any good statistical analyses. Well, there is this article by Plassman and Tideman which is pretty good and looks the Lott data on crimes and shall issue CCW laws. Nope, don't bother looking for that article. Ignore this article at the Statistical Assessment Service (a.k.a. STATS). That article is pretty well balanced that points out that there is bogus numbers on both sides.

Kevin then links to this website which is, I guess, his idea of good data analysis. Unfortunately the website starts out very badly. Tim Lambert puts forward his view of Lott's argument:

1. that there were more guns
2. that there was less crime
3. that more guns caused less crime

Lott's argument depends on all three parts being true. If any one of the parts is incorrect, the entire argument fails.

The problem is that number one does not have to be true for the conclusion to be valid. You don't need more guns, just more people carrying (concealed) existing guns legally. The idea is that when this occurs then criminals will percieve a higher level of risk for the same payoff and will look elsewhere for a payoff (say stealing cars). In fact, Tim Lambert is being extremely dishonest, IMO, with number three in that list; while Lott's title is indeed More Guns, Less Crime it should not be taken literally (or as Mr. Lambert's mother undoubtedly told him, don't judge a book by its cover...or in this case simply its title). Lott himself notes that there is a substitution effect when shall issue CCW laws are put into place. Criminals substitute away from crimes involving other people (now with a higher probability of being armed) to crimes that do not involve people such as stealing cars and burglarizing homes when people are not in the home.

Lambert also has this comment here

In a critique of econometric studies such as Lott's[15] Ted Goertzel makes an important point--for any study that involves a multiple regression that finds a significant association, it seems that there is another study that applies a different model to the same data and gets a different answer. There are several examples of this happening with Lott's study below.

Goertzel argues convincingly that

When presented with an econometric model, consumers should insist on evidence that it can predict trends in data other than the data used to create it. Models that fail this test are junk science, no matter how complex the analysis.

In the case of Lott's model we are in the fortunate position of being able to test its predictive power. Lott's original data set ended in 1992. Between 1992 and 1996, 14 more jurisdictions (13 states and Philadelphia) adopted carry laws. We can test the predictive power of Lott's model by seeing if it finds less crime in those jurisdictions. Ayres and Donahue [2] have done this test. They found that, using Lott's model, in those jurisdictions carry laws were associated with more crime in all crime categories . Lott's model fails the predictive test.

Ayres and Donahue go on to examine all the states adopting carry laws using data up to 1997 and found that carry laws were associated with crime increases in more states than they were associated with decreases. They rather pointedly observe that

Those who were swayed by the statistical evidence previously offered by Lott and Mustard to believe the more guns, less crime hypothesis should now be more strongly inclined to accept the even stronger statistical evidence suggesting the crime- inducing effect of shall issue laws.

There is only one problem here. If the model is incorrectly specified as this argument claims then relying on said model's results is not any better when the results are what you want them to be or your opponent wants them to be. The solution isn't to say, "See, now it says we're right and you're wrong, neener neener neener," but to try and correct for the mis-specification of the model. Again, this is why statistics is, IMO, rocket science. Even smart people get easily confused when dealing with statistics.

Lambert also mischaracterizes the findings of the Plassman and Tideman analysis linked above.

Plassmann and Tideman [39] point out that Lott's analysis technique assumes that crime rates are normally distributed and that this is not even close to being true for low crime counties. When they made some plausible changes to the specification, the effects on murder vanished. However, when they did their own analysis assuming that the murder rate was Poisson distributed, they found an even stronger effect (a 12% decrease). They also looked at the effects on each state and found a confusing pattern of results, with the effect varying from a statistically significant increase of 6.5% (Virginia) to a statistically significant decrease of 35% (Montana). While we would not expect the laws to have exactly the same effect in every state, it seems hard to see how the effects could be so radically different.

Their conclusions when they included the state effects was that shall issue CCW laws should still be considered as policy responses to crime, but that they shouldn't be considered blanket solutions (i.e. whether or not a state enacted them should be carefully considered because not in all states was there a reduction in various violent crimes).

Finally, the idea that because Lott's work might be invalid (bad statistical methodologies, etc.) to conclude that the overall conclusion is wrong is an example of the fallacy fallacy. Lets suppose the model is incorrectly specified. Suppose a relevant variable (or several variables) are missing. Does it then follow that the conclusion that shall issue CCW laws result in higher crime rates? No. All that you can say is Lott's models don't work or need refining, etc. In other words, Kevin's link is far from high quality. Reading some of the literature cited in that link might be a better strategy to trying to get some insights into this argument. However, the gun/crime issue is a complex one. For example, shall issue CCW laws might increase gun ownership rates, which could mean that more guns are stolen from households during burglaries. These illegal guns might end up in the hands of criminals thus increasing gun crimes. Disentangling this omlette of data is very challenging. Lott's work, is IMO, an important first step.

Update: Clayton Cramer has a post on this. At the bottom he notes that Lott has done a second survey that supposedly supports the results of his first survey. Kevin makes this claim about this new information:

This completely evades the point. No one is all that concerned with the precise results of the survey — which are not especially crucial to his overall thesis — but with the fact that he apparently lied about conducting a survey in the first place.

Of course, lets ignore the fact, that if using the same approach Lott gets the same result (or nearly the same) then it does indicate he did the initial study.

Update II: Norhtwestern University Law Professor James Lindgren is taking this very seriously and is giving this issue a very, very thorough look. You can see his progress here.

Update III: Testing Lott's Model:

Up above I note the sloppiness of the Lambert/Goertzel attack on Lott's model. The argument is that Lott's model is flawed because when more data is added the parameter values changed for shall issue CCW's. However, the change in parameter values over time and as more cross sectional data is added isn't unheard of. To me a better test would be to take Lott's estimated parameter values, take the new data for the variables and see what the model says for various crimes. If these match up to what is observed then Lott's model is a good predictor, if not then it is not a good predictor and should be reworked, junked or ignored. Re-fitting a model that Lambert et. al. are working so hard to show is flawed then claiming the result, results that Lambert et. al. now like, is bad statistics.

Update IV: More on Ted Goertzel

Ted Goertzel is one of the guys that Tim Lambert uses in his efforts to debunk John Lott. But now we can see some of the hypocrisy of Tim Lambert by relying on Goertzel. Goertzel writes the following:

The problem with this, as anyone who has had a course in statistics knows, is that correlation is not causation. Correlations between two variables are often "spurious" because they are caused by some third variable. Econometric modelers try to overcome this problem by including all the relevant variables in their analyses, using a statistical technique called "multiple regression." If one had perfect measures of all the causal variables, this would work. But the data are never good enough. Repeated efforts to use multiple regression to achieve definitive answers to public policy questions have failed.

Prof. Goertzel is right to a degree. Correlation is not cuasation, that is (or should be) hammered home in the very first statistics class anybody takes. He is also correct about spurious correlation and omitted variables. The problem is that we live in the real world and while the price of oranges in Oregon might have some ralationship to the price of Hyudai's in Tampa, Florida (assuming a general equilibrium model) most researchers are not going to bother including said variables. Econometric modelling is not about literal reality. To make a model that reflects literal reality is an oxymoron. Models are, in the words of Edward Leamer,

A model is a powerful device for organizing our thoughts; it is not literally true: indeed it derives its power from the very fact that it is not literally true. Thus there is no reason to test it. Instead of testing, we should be determining its accuracy and usefulness. We should attempt to identify empirical circumstances in which the model is useful and other circumstances in which it is misleading. (emphasis in the original)

So the tests aren't whether or not some correlations look funky (as Lambert argues) or what the t-test statistics are (as Lambert does), but whether or not the model gives the right answer, i.e. does it predict well. Taking the new data that is available and seeing if it predicts the correct crime rates. If it doesn't then there is a problem, if it does, then the model is useful. Nobody, that I know of has done this with Lott's model. We have people complaining about measures of goodness of fit, spurious correlations, and re-fitting the model.

Further, Goertzel's criticism is sweeping and damning not just of Lott's analysis but of ALL forms to statistical analysis and not just on the gun debate, but wherever statistical analyses are performed. The problem is that statisical models can be good predictors, the real problem are the people who have access to softward, a little bit of statisical training and too much time on their hands to paly with their statistics programs. Goertzel's criticism applies not only to Lott, but to all the other studies that Lambert lists on his site. Yet, we don't see the same standards being applied to these studies that Lambert applies to Lott.

Goertzel goes on:

Lott's work is an example of statistical one-upmanship. He has more data and a more complex analysis than anyone else studying the topic. He demands that anyone who wants to challenge his arguments become immersed in a very complex statistical debate, based on computations so difficult that they cannot be done with ordinary desktop computers.

Maybe Prof. Goertzel needs to meet with his Department Chair about a new computer or something. I myself have managed to plough through gigabytes of data with my old piece of junk desktop PC. It was a 233 Mhz, machine with 126 meg of RAM and using Windows 95. Somehow, I managed to do it. Second, the gun issue is not going to be settled by simple analysis. It is a complex issue and is going to require complex analysis.

Goertzel continues with this as well:

He challenges anyone who disagrees with him to download his data set and redo his calculations, but most social scientists do not think it worth their while to replicate studies using methods that have repeatedly failed. Most gun control researchers simply brushed off Lott and Mustard's claims and went on with their work. Two highly respected criminal justice researchers, Frank Zimring and Gordon Hawkins (1997) wrote an article explaining that:

just as Messrs. Lott and Mustard can, with one model of the determinants of homicide, produce statistical residuals suggesting that 'shall issue' laws reduce homicide, we expect that a determined econometrician can produce a treatment of the same historical periods with different models and opposite effects. Econometric modeling is a double-edged sword in its capacity to facilitate statistical findings to warm the hearts of true believers of any stripe.

Zimring and Hawkins were right. Within a year, two determined econometricians, Dan Black and Daniel Nagin (1998) published a study showing that if they changed the statistical model a little bit, or applied it to different segments of the data, Lott and Mustard's findings disappeared. Black and Nagin found that when Florida was removed from the sample there was "no detectable impact of the right-to-carry laws on the rate of murder and rape." They concluded that "inference based on the Lott and Mustard model is inappropriate, and their results cannot be used responsibly to formulate public policy."

There are numerous problems here. First, is this notion that nobody should take Lott's data and do any analysis (I seriously doubt Lott was implying people should duplicate his analysis...the numbers will undoubtedly be the same) is the hallmark of a junk scientists. Second, the Black and Nagin study isn't flawless either. The Plassman and Tideman analysis above is actually better in that it has a better way of correcting for some of the problems noted by Black and Nagin. Of course, Goertzel doesn't list this study and Lambert mis-states the conclusions.

Overall, this paper is not something Lambert should, ideally, point too, but when on a mission finding anything to throw at your opponent is what you do, I guess.

Update V: Tim Lambert, despite his obvious bias against Lott, has probably the best site for information on this potential scandal. To keep informed on this I recommend that site and specifically the page with Lindgren's progress.
Steve 10:36 PM