Semi-Daily Journal

Well I have to agree on “Dow 36,000.” Glassman and Hassett’s book is a joke. What’s more they aren’t even repentant. Here we can be specific. They misrepresent Siegel on stocks being no more risky than bonds (Siegel himself says so). They don’t seem to understand the Gordon dividend discount model. Then they use earnings growth instead of dividend growth in making calculations. Unfortunately for G and H they made near term predictions unlike (say) Paul Ehrlich. If AEI should fire G and H for making silly predictions, should Stanford fire Ehrlich?

Well, the Washington Post still pays James Glassman to provide investment advice.

Ehrlich made a whole bunch of faulty predictions: mass famine in the Indian subcontinent by the end of the 1970s comes to mind.

Brad DeLong writes:
> Ehrlich made a whole bunch of faulty predictions: mass
> famine in the Indian subcontinent by the end of the
> 1970s comes to mind.

Ehrlich would apparently say all kinds of outrageous things; England not existing in 2000 even better than what you say. After he got his clock cleaned in the bet with Julian Simon, I detect a bit more rationality in his predictions. So one prediction he offered a Simon a bet on was that world wheat production would decrease from 1994 to 2004. Interestingly, that might turn out to be right. But it wouldn't necessarily mean what Ehrlich wants it to mean; the economics of commodities still suck, and markets are probably freer now than then even in agricultural goods.

Brad, I think you or Krugman or somebody should carefully document this meme that Murray, and apparently Herrnstein too, "don't know what a correlation coefficient is". If you just continue to quote it without explicit support, you're just as bas as you claim they are.

"Ehrlich made a whole bunch of faulty predictions: mass famine in the Indian subcontinent by the end of the 1970s comes to mind."

"Astrology" does not interest me, but even though India has no food shortage and should have none, there is woeful food distribution through India. There is fierce hunger and malnutrition, and no productive reason for either.

Distribution always distribution. Think of how readily available medicine can be and how many millions of African lives are being needlessly lost for lack of treatment of AIDS.

Notice that the Administration often chooses to release policy reports at American Enterprise. Yuch.

Does anyone pay attention to what Paul Ehrlich says anymore? Is he still making predictions? On topic, does anyone pay attention to what the AEI says anymore except it's ditto heads?
However, in the 70's there were others who were certain that mass famine could happen in India. I remember a speaker at work, an agronomist, who was very emotional when speaking of this possibility. A failed monsoon would do it.

Again, I wish to emphasize that hunger is now a fierce problem in several countries including India. The problem in India is not production but distribution, while in southern Africa the problem is production - environmental, medical, foreign resource depletion - and at times distribution.

AIDS is a severe threat to food production in parts of southern Africa.

I wonder -- can all those smart people criticizing Murray's statistical acumen have been wrong? a canonical example of the critiques I found may be read at http://www.skeptic.com/03.3.fm-sternberg-interview.html

Sternberg -- as he points out, no left-wing critic of intelligence testing itself -- relates, among other things:

"To draw causal inferences from correlational data, which is what all their data are, is statistically incorrect."

"[in comparing correlations] they don't take into account the reliability and precision of the measures being used."

read the whole interview; agree or disagree with Sternberg's point of view, you'll likely find it both interesting an persuasive.

WCW I think you have the wrong comment page, but that’s ok. I read the Sternberg interview with the Skeptic, and I don’t see any real criticism of M&H. Sternberg has his 3 factor model, and he thinks it does better (in some sense) than the conventional general intelligence that M&H use. But in that interview Sternberg does not provide any specific criticism of M&K methods. Nor does he say they used invalid data. I’m not an advocate for M&H, but I want to know where they went wrong in their methods. Note that Sternberg says that M&H state a range for IQ heritability, 40% to 80%. Some here are trying to tell us that M&H assert 100%.

"Ehrlich made a whole bunch of faulty predictions: mass famine in the Indian subcontinent by the end of the 1970s comes to mind."

Paul Ehrlich circa 1975: "I would take even money that England will not exist in 2000."

What an imbecile (to use his own word, regarding Julian Simon).

And they gave the Ehrlich a MacArthur "Genius" Award.

"So one prediction he offered a Simon a bet on was that world wheat production would decrease from 1994 to 2004. Interestingly, that might turn out to be right."

Based on information presented on this site, I don't think so:

http://www.spectrumcommodities.com/education/commodity/w.html

"World wheat production in 1999/2000 of 578 million metric tonnes compares with 588 million in 1998/99 and the record high 1997/98 crop of 609 million tonnes. Production in the mid-1990's averaged about 540 million tonnes."

The increase from 1994 to 2004 may not be substantial (maybe less than 10 percent) but there will probably be an increase.

"But it wouldn't necessarily mean what Ehrlich wants it to mean; the economics of commodities still suck, and markets are probably freer now than then even in agricultural goods."

According to that website:

http://www.spectrumcommodities.com/education/commodity/w.html

"The 1999/00 average price received by farmers was forecast to range from $2.45 to $2.65 a bushel vs. $2.65 in 1998/99 and the 1990's high of $4.55 in 1995/96."

So the prices fell from $4.55 a bushel in 1995/96 to $2.45 to $2.65 in 1998/99. That's gotta hurt, if one is a wheat farmer! Good for consumers, though.

apologiae to any others still reading this long-dead comment thread for what follows.

to Mr. Zarkov:

one, in re 'think you have the wrong comment page', please see the root post for these comments, 'Charles "I Don't Know What a Correlation Is" Murray' and Dick Thompson's comment to which I responded. I have the right thread.

two, in re 'I don’t see any real criticism of M&H', I'll pick just one quote. Sternberg says, "having committed themselves to a position, if the data are consistent with their position, they cite them, and if the data are inconsistent, they try to come up with what I think are fairly hare-brained interpretations of how that could happen." the piece contains heaps of real criticism of M&H.

three, in re 'Sternberg does not provide any specific criticism of M&K methods', Sternberg recites a litany of specific criticisms, from those I quoted in my initial comment, to criticisms of the main study they used, to others of how they distorted another study. the interview is rife with specific criticism.

four, in re 'nor does he say they used invalid data', viz Sternberg: "virtually all their data are based on one study, the National Longitudinal Study of Youth (NLSY), which was not a study that was particularly representative of the United States population." there are other critiques as well.

I hope it's comfy down there between the hardwood and the shoeleather.

Beyond the no-nothing use of statistics, "Bell Curve" was written to attack African-Americans and civil rights extensions. This was a bad faith book.

The Lott/Ehrlich comparison is interesting. Leftists continue to savage Lott for his actions but Ehrlich continues to receive praise. There was a nauseatingly glowing profile of Ehrlich a few years ago in Scientific American. If countries had followed Ehrlich's advice, millions would have died horrible deaths.

Murray, Glassman, Lott . . . Don't forget Nebish D'Souza. And Newt.

WCW you are absolutely right. I missed paragraphs in the center of the article where Sternberg certainly does make specific criticisms of M&H’s methods. I agree that measurement error will degrade correlations, and that’s definitely a defect in M&H. The limitation of “Bell Curve” to one non-representative data set (if that’s true) is another problem. Induced correlations from selection bias is yet another good point.

In a few cases Sternberg’s prose is a little muddled if not contradictory, perhaps as the result of transcription errors. For example:

Sternberg: “… Herrnstein and Murray invite the reader to conclude that race differences are due to genetics …”

Yet in the next paragraph:

Sternberg: “There is nothing in the book that suggests that race differences are genetic. They even say that.”

When Skeptic asks Sternberg to name studies on race difference that M&H left out, Sternberg won’t answer, he responds by saying what they do cite, albeit with criticism.

Skeptic: Which studies don't Herrnstein and Murray cite?
Sternberg: Well, one study that they cite and distort the results of is the Scarr-Weinberg study.

Finally we come to the matter of correlation and causation. The M&H analysis appears to use mainly if not solely, observational (happenstance) data and not designed experiments to advance the central thesis of the book. They say that American society is becoming more stratified because IQ is more able to drive success than in the past. They then present correlations (some spurious) to support their thesis. To the extent that M&H have done defective modeling (and it appears they have) their conclusions lack a firm foundation of evidence. This is a valid criticism. But bear in mind that observational data is used all over the place to draw conclusions and take actions. For example someone will subpoena salary data from a company and do a multiple linear regression: salary= b_0 + b_1 x_1 + b_2 x_2+ … b_N x_N + b_N+1 sex. And N can be large, like 8. When they find b_N+1 has a large t-value, a sex discrimination suit gets launched. The lawyers do think correlation is causation. But high dimensionality here dooms the analysis (see the work by Leo Breiman).

Erlich picked five metals in 1980,and bet their prices would be higher in 1990. Maybe he should have talked to a geologist and an economist. Hope he was smarter with his retirement funds.

A while back Simon wrote a paper on endangered species with Aaron Wildavsky that matched Erlich's weird ideas on population.

I've never understood the traction either of these guys got; guess there is always going to be a market for snake oil.

Ehrlich didn't understand the basic economic truth of commodities - and that is that it is extremely inflexible capital. One cannot stop mining copper and start mining gold, if there is no gold to be had. Commodities production went up relentlessly for a long period simply because one commodity: oil, was undervalued.

====Brad, I think you or Krugman or somebody should carefully document this meme that Murray, and apparently Herrnstein too, "don't know what a correlation coefficient is". If you just continue to quote it without explicit support, you're just as bas as you claim they are.

As if it hasn't been done. My particular favorite is the chapter in Wilson that debunks the majority of Losing Ground in like 20 pages.

Of course, the best documentation is Losing Ground and the Bell Curve....

Murray is simply the public face of a conspiracy by untenured Sociologists to create a bunch of bad arguments they can attack and get tenure from doing so.