What Happened in Iraq in August?

=======================

UPDATE: if you are arriving here from the NRO Corner, you may wish to click here after reading the post below. My updated post is a response to an e-mail message that Rich Lowry posted on the Corner (which took slight issue with my analysis).

=======================

In an earlier post, I warned against the tendency to overinterpret random error. In July, civilian casualties in Iraq were (randomly) high, and much was made of that fact. Even the stalwart Max Boot began to fret:

The security situation in Baghdad has been in free fall since the bombing of the Golden Mosque in Samarra on Feb. 22. In retrospect, that attack appears to be a turning point when the chief problem in Iraq went from being a Sunni-dominated insurgency to a civil war in which Shiite and Sunni militias are equally culpable. The result has been a horrifying surge in violence, with about 100 Iraqis dying every day, the bulk of them in Baghdad.

That's the kind of impression one gets from reading hysterical media reports. The bombing of the Golden Mosque was, indeed, an unfortunate turning point, as I explain below, but the idea that the security situation is a free fall is a hyperbolic and unnecessarily demoralizing way to frame it. I'll show some charts to make that clear. Here is my 2-point conclusion in a nutshell:

1. Sectarian violence aimed at unarmed Iraqi civilians (what we should call sectarian terrorism) immediately jumped to a higher, but stable, level after the bombing of the Golden Mosque

2. The insurgency fighting U.S. and Iraqi forces is slowly but steadily weakening

It's hard to believe this when the media knowingly proclaims that the insurgency is reaching new levels of strength and that sectarian violence in July spiraled to unprecedented levels. A few weeks ago, the New York Times concluded that:

The number of roadside bombs planted in Iraq rose in July to the highest monthly total of the war, offering more evidence that the anti-American insurgency has continued to strengthen despite the killing of the terrorist leader Abu Musab al-Zarqawi.

Nonsense. All signs indicate a slowly weakening insurgency despite the roadside bomb statistic (which is the only statistic that seems to suggest otherwise). And the BBC had this to say:

Officials at Baghdad mortuary say they received 1,855 bodies in July, as the capital remains gripped in a wave of violence which has beset it for months. The figure - the highest yet - is a rise of more than 350 on June. Officials say about 90% of the deaths were as a result of violence.

The impression created by that kind of wording greatly differs from a more sober look at the evidence.

Measuring casualties in Iraq is problematic because people keep changing the method. When they do, you don't know if the new result represents a real change or if it just indicates that a different method is being used. The methods used by the media hop all over the place, but Iraq Coalition Casualties has been using the same method (tabulating deaths reported by the media) for a long time. They do not detect every death by any means, but their method is very definitely sensitive to change, and the question of most interest right now is how things are changing in Iraq.

In an earlier post, I showed the data for civilian casualties in the 5 months following the mosque bombing (March through July), and I also included a predicted value for August based on the average of those 5 months. Here it is again (the predicted value is in blue):

The predicted value for August was 897 killed, assuming things were stable. That would not be a heartening figure, but it would counter the notion that the security situation is in an uncontrollable free fall. The actual value for August is now known, and here it is (again, it's the blue bar):

The actual value turned out to be 949 (updated to 970 since I made the chart), which is only a bit above what one would expect. In fact, the two charts look almost identical. As you can see from this figure, violence is not escalating relentlessly since the bombing of the mosque. Instead, it has been holding steady over a 6-month period. I wish the news were better (i.e., I wish civilian deaths were clearly declining), but I am happy that the situation is not ever worsening, as many seem to believe it is. It could be a lot worse than it is, and there is no need to create the impression that it is.

Now let's consider some longer trends. To average out the noise a bit, I looked at 3-month blocks (instead of looking at month-to-month values) in the next chart:

The two red bars show the 6 months since the bombing of the mosque (each red bar represents the average value over a 3-month period). The 3 blue bars show the 9 months preceding that event (3 months at a time). As you can see, sectarian violence against civilians immediately jumped to a new but stable level after the February bombing. That supports the first conclusion I mentioned above.

I made some similar charts to underscore my further claim that the insurgency (which is distinct from the sectarian violence) is gradually weakening. First, I plotted the data for U.S. military casualties in 3-month blocks:

U.S. casualties are, if anything, decreasing, which I take to be a probable sign of a weakening insurgency. Then again, one might imagine that this decrease simply reflects the increasing security responsibilities that are being assumed by Iraqi security forces. To see if that might be true (i.e., to see if their casualties are on the rise commensurate with the drop in U.S. casualties), I next looked at Iraq military/police casualties in the same way:

They, too, are lower than they were before.

The upshot? Sectarian terrorism is at a new and stable level, and the insurgency is gradually weakening. That's what appears to be happening in Iraq. Whether you support the war or oppose it, it's best to let the evidence (instead of hyperbolic media claims) influence your thinking.

Is our Unhappiness Caused by "Stagnant Wages?"

If there is one theory that people are resorting to in order to explain our dour economic mood in spite of all the good economic data, it is that wages have not kept up with inflation. Even Tigerhawk uncritically buys into this theory (via Instapundit). It's a great theory, and it dovetails perfectly with the "Two Americas" speech that John Edwards used to give. I love that speech. I'd love it more if it were true. But it isn't, so far as I can tell.

I don't doubt the wage statistic, but I do question the reflexive interpretation of it (i.e., that many people are actually worse off). Don't let the media do your thinking for you. Doubt, then verify.

It stands to reason that employers will not raise wages for the fun of it. They raise wages when they have to. When there are qualified people in the workforce who you can hire on the cheap, employers will tend to do that instead of raising wages. But when more people who want to work actually are working, people are better off on the whole even if wages remain the same. And more people are working now.

Once the unemployment rate reaches a minimum (thought by some to be about 4%), upward pressure on wages becomes strong. When that happens, you can be sure that the media will say that because wages are now on the rise, inflation will rear it's ugly head (i.e., that will be the newsworthy issue). Look at this report from back in February in Businessweek:

U.S.: Low Unemployment Raises An Old Inflation Debate

Vigorous job growth, including 687,000 new payroll slots in just the past three months, pushed the jobless rate down to a 4 1/2-year low of 4.7% last month. Hourly pay of production workers, up 3.3% from a year ago, is rising at the fastest pace in nearly three years. And job growth is generating plenty of income to support consumer spending, which is a key reason for the economy's forward thrust in early 2006. That momentum will only tighten the labor markets further.

OK, but my real point is that a better gauge of how people are doing is to look at median family income, not median wages. I have only found these numbers through 2003, and here they are (charted, of course):

The data are plotted by quintile, covering people in the lowest 5th of the population (i.e., the poorest people), to the upper 5th. The top 5% are also included so we can see how the really rich are faring. The blue bars represent the average from 1982 to 1992. The magenta bars represent the Clinton Years and the yellow bars represent the Bush years (excluding the recovery years of 2004 through 2006, where the numbers are sure to look even better). The income figures are in constant 2003 dollars.

What you see is that no one is worse off, but the richer you are, the better off you are (especially during the Clinton years, during which the top 5% made spectacularly impressive gains). When I can find the numbers for the recovery years, I will be sure to present them. For now, my point is that during the Clinton years, we loved the economy. Now we don't. This is true even though the median-family-income story was much the same during the two eras (though it is slightly better now, overall, than it was then). If we hate our fabulous economy now because the less fortunate are not gaining much ground, we should have hated it back then, too.

In a way, though, there are two Americas, but that has been true for a very long time (i.e., it is not unique to the Bush years). The poorest among us are not gaining much ground. They didn't during the Clinton years and they aren't during the Bush years. So maybe I can continue to love that speech by John Edwards after all. A matter that needs to be researched is the effect of immigration on the lowest quintile. I suspect that their wages and their incomes would both be increasing were it not for illegal immigration. I'm not suggesting that we seal the borders, necessarily. I'm just pointing out a factor in the consistent lack of income growth for the lowest 5th of the population. The effect of illegal immigration on that group is worth thinking about.

Fretting over the Federal Deficit

In my last post, I graphed up some poll results showing that we are rather displeased with what is actually a fabulous economy. The economic figures I presented were all quite positive, but there are a few warts on the economy that might help to explain our dour economic mood. One possibility is the size of the federal deficit. The way this issue is covered in the press is sure to demoralize anyone. The problem appears to be that reporters lack basic knowledge, so they are inclined to do dopey things like report the absolute size of the deficit instead of reporting the deficit as a percentage of GDP. The absolute size of the deficit does not tell you very much, just as the absolute size of someone's credit card debt is not that informative. Running up $10,000 of debt in a year might sound like a lot, and it is if your salary is only $20,000. But it's a drop in the bucket if your make $2,000,000 per year. What matters is the size of your spending deficit relative to the size of the your personal economy. The same principle applies to the federal deficit. Even so, when reporters write about the deficit, they almost always focus exclusively on its absolute size. Here is a typical example:

Record '05 Deficit Forecast
War Costs to Raise Total to $427 Billion

By Jonathan Weisman
Washington Post Staff Writer
Wednesday, January 26, 2005; Page A01

Additional war spending this year will push the federal deficit to a record $427 billion for fiscal 2005, effectively thwarting President Bush's pledge to begin stanching the flow of government red ink, according to new administration budget forecasts unveiled yesterday.

Well, that ought to put a damper on your feelings about the economy! Nary a word in this article about the size of the deficit relative to the GDP. Based on articles like these, many people probably think that 2005 was the worst year ever in this regard. But they'd be way wrong.

Let's take a look at the federal deficit as a percentage of the GDP, which is the only sensible way to do it, and let's consider some historical context as well (as every news article about the deficit should):

Over the years, we've gone from a record deficit of 6% of GDP (in 1983 under Reagan) to a record surplus of 2.4% (in 2000 under Clinton). In 2005, the year that had the attention of that hysterical Washington Post reporter, the deficit was historically typical (about 2.6% of GDP). But that's not what Jonathan Weisman told his readers. He said "Additional war spending this year will push the federal deficit to a record $427 billion for fiscal 2005." 2005 was not even close to a record in proper terms (i.e., in terms of the GDP), and 2006 is projected to come in even lower, at about 2.3%. Sheesh!

The last 6 points on the plot show the Bush years. Things were in free fall in years 2001, 2002 and 2003 after reaching a nice surplus in 2000 (Clinton's last year in office). As I noted before, these are the years that were affected by both the dot-com crash and the 9/11 attacks (not to mention the wars in Afghanistan and Iraq). As the economy has rebounded, the deficit numbers have returned to typical levels, and, as you can plainly see, they are nowhere near the record-setting deficits that reporters love to fret about.

It could be that the return of deficit spending, coupled with news reports that focus solely on the absolute size of the deficit (which leaves readers with the impression that the deficit is unprecedented) accounts for the sour mood. Is it too much ask that reporters who write about the federal deficit do so in a way that clarifies rather than misleads? Probably. But if they did, I suspect that the next Gallup poll on the economy would show Americans to be in a better mood.

By the way, before you give Clinton too much credit for those surplus years, take note of the fact that they preceded the stock market crash. Here is anti-Bush economist (and New York Times columnist) Paul Krugman weighing in on that:

[Question]: And implicit in that is the notion that the Clinton economy was phony -- it was just the dot-com bubble. Is that a fair criticism?

[Paul Krugman]: No. There was a bubble. The one piece of it is to say that the budget surplus was inflated by the stock market bubble, that capital gains were certainly giving us a lot more revenue.

The huge budget surpluses of the Clinton Administration had a lot to do with the stock market bubble. They were not really a lasting achievement.

The gradual return to a balanced budget was a real achievement. But that move into surplus was the stock bubble.

I am not knocking Clinton. I'm just trying to provide some perspective for those who take pleasure in knocking Bush.

Even after you exclude the effects of the bubble, it is fair to say that the deficit situation under Bush is worse than it was under Clinton, but the media narrative (surpluses under Clinton, followed by record deficits under Bush) may be sufficiently misleading to explain why Americans mysteriously hate their fabulous economy.

UPDATE: Speaking of reporters trying to demoralize their readership, this headline is a classic:

GDP growth revised up less than expected

By Mark Felsenthal Wed Aug 30, 12:40 PM ET

WASHINGTON (Reuters) - The economy grew at a faster clip in the second quarter and inflation was slower than originally reported, a government report said on Wednesday, suggesting the Federal Reserve will keep interest rates steady.

The U.S. economy grew at a 2.9 percent annual pace in the second quarter, faster than the 2.5 percent rate initially reported but marginally below what analysts were expecting, as higher business investment offset a drop in residential construction, a Commerce Department report showed.

You have think for a moment to realize that there is good news about the GDP in this article and separate news about the expectations of economists. The way the good news is presented is not likely to improve the way Americans feel about the economy.

Americans Hate their Fabulous Economy

Polls consistently show that Americans are rather negative about the economy, and many have wondered why that might be considering that our economy is the envy of the world (as well it should be). The source of our pessimism is a bit of a mystery, and it is one that I am going to pursue. For the moment, I'll simply note that I have identified what may be a contributing factor to our unwarranted despair: reporters rarely use charts, and that allows them to wax poetic about the "jobless recovery," "worrisome signs of inflation," "record setting deficits," and other such nonsense. Charts anchor the mind to the raw data, which makes it hard to use bumper-sticker slogans that mislead (and demoralize) more than they clarify. In any story about the economy, there is never an excuse not to use a chart, but reporters almost never do.

First, I'll show you a chart that illustrates how pessimistic you are these days compared to earlier years. To make this chart, I went to a well known polling site, copied the numbers into an Excel spreadsheet, and presto:














So, in 1998 (during the Clinton era), Americans were 65/35 on the optimistic side, but now (during the Bush era) we are 65/35 on the pessimistic side. A complete reversal. Are things really that bad? No. In fact, the evident pessimism is borderline ridiculous, but it's hard to appreciate that unless you look at some simple charts. I'll be looking at quite a few as time goes on, but let's consider some right now.

Growth of the Gross Domestic Product (GDP) is, perhaps, the most basic measure of the health of any economy. A growth of 1% is often described as anemic, 2% is so so, 3% is good and 4% is excellent. Let's take a look at that measure during the Clinton and Bush eras (I got these numbers here):

Give the man credit: GDP growth was very good under Bill Clinton (averaging 3.7%). It took a hit during the early years of the Bush administration, but it has snapped back smartly since, right back to where it was during the Clinton years (the 2006 value is an average of the first 2 quarters of the year). Why did things go south for a while? First, the stock market bubble popped. The Nasdaq's long slide (i.e., the dot-com crash) began in March of 2000 and continued until October of 2002. That sucked a lot of life out of the economy. Second, September 11, 2001 happened. In light of those two unforseen calamities, it's amazing how well the economy has been doing in 2004, 2005 and so far in 2006. Whether or not Bush's policies are responsible for the impressive recovery is a matter of debate, but it is clear that, at a minimum, his policies have not choked off economic growth. Give that man credit, too. Except no one does. Despite the fact that GDP growth is right back to where it was during the Clinton era, Americans don't feel the same way about the economy, to say the least. Why not?

The jobless recovery, perhaps? No story about the so-called jobless recovery should have gone to print without some basic charts that would provide some perspective. Here's a useful, perspective-supplying chart (data acquired here):

What you can see is that unemployment has been relatively low all throughout the Clinton and Bush administrations. Thus, all of the obsessing about the "high" unemployment rate during the last presidential campaign was so much nonsense. It was made possible because reporters avoided using charts like this one. Let's zoom in on the unemployment situation just during the Clinton and Bush years:

The economy bounced back in terms of GDP by 2004, then the unemployment rate -- which was never very high -- began to drop as well. In fact, it's now right back to where it was during the Clinton years. But back then, we liked the economy. Now we don't.

Inflation? Is that the problem? Let's look at the chart (based on data obtained here):

Inflation during the Bush administration has been much like it was during the Clinton administration. Even so, back then, we liked the economy. Now we hate it. So, what exactly is the problem? The "record setting" budget deficits, perhaps? Not really. Stagnant wages? Maybe, but I doubt it. I'll take a look at these a bit later, but for now, my point is that any story you read about some aspect of the economy ought to include simple charts like these. Those two stories about budget deficits and stagnant wages -- like almost all stories about the state of economy -- don't do that. You can learn more from a few informative charts than you can from reading the words of a reporter who has an agenda that is advanced, not by showing you the actual numbers, but by using bumper-sticker slogans to create the impression that things are "spiraling out of control." Oh wait, that's the phrase reporters use to characterize Iraq. Well, they don't use charts for that purpose, either (and for the same reason).

In Search of a Serious Democrat on Iraq

Other than Joe Lieberman, that is.

In her latest NEWSWEEK column, Eleanor Clift addresses a question I asked in a post a few days ago. In that post, I asked:

Isn't it fair to describe the two political options we face [with respect to Iraq] as follows?

Option A ("phased withdrawal"): Let the unrestrained slaughter of innocent civilians in Iraq begin.

Option B ("stay the course"): Keep the troops in Iraq until they can facilitate the political climate and train the Iraqi security forces needed to keep that slaughter from ever happening.

Clift, who supports withdrawal, takes Bush to task for making mention of option A. Her column is entitled "The Fear Factor: President Bush’s new argument for staying the course in Iraq: It’ll only get worse if we leave." Well, I think it will get worse, too, and that's why I believe that humanitarian concern for Iraqi civilians translates into support for Bush's "stay the course" alternative. So I was excited to see a proponent of troop withdrawal address this issue head on. I could hardly wait to see what she had to say. Here is the substantive paragraph from her column:

If you think things are bad now, they will be worse if we leave. That’s the essence of President Bush’s argument for staying the course in Iraq. Bush is doing what he always does—shamelessly ramping up the fear factor. He says if U.S. troops leave Iraq, the terrorists will be right behind them, bringing Baghdad to America. He’s brought ruin to Iraq and his policies are helping create our worst nightmare, a nuclear Iran. How much worse can it get?

How much worse can it get? Well, how about 5 times worse, for example? Instead of 3000 civilians killed per month, we could have 15,000 killed per month. Is 3000 an agreed-upon upper limit on the number of civilians that sectarian militias and terrorists will kill in any given month? Is there some analysis somewhere suggesting that Iraq has reached maximum entropy, such that the chaos can no longer expand? She doesn't say, which means she simply assumes the problem away. A problem that virtually everyone else -- including our military commanders on the ground in Iraq -- agree exists. That seems disingenuous to me.

Although I am looking for a serious response to my question (i.e., isn't "phased withdrawal" tantamount to "let the unrestrained slaughter begin?"), it looks like I'm not going to get it. The Washington Post recently surveyed the campaigns of 59 Democrats in hotly contested House and Senate races. Here's what they found:

...a majority agree with the Bush administration that it would be unwise to set a specific schedule for troop withdrawal, and only a few are calling for substantial troop reductions to begin this year, according to a Washington Post survey of the campaigns.

and

With polls showing that a majority of Americans believe it was a mistake for the United States to invade Iraq, some Democrats say the wisest political course is to blame Bush and the GOP for problems in Iraq but avoid getting drawn into a debate with Republicans over how they would go about dealing with the war.

I can scarcely believe that these Democrats would get a single vote, much less enough votes to actually win. They want to stand on the sidelines, carping at their own president in the middle of war, without saying a word about they would do to address the problem. That's not serious, and these are serious times. If they oppose the war, fine, but they have to say precisely what they'd do. Instead, they are choosing to act like adolescents.

I'd like to vote for Democrats again, but I'm not going to do it until they get serious.

The Valerie Plame Fable is a Psychological Mirror

It is, of course, extremely well-documented by now that Bush did not lie about Iraq's WMDs, but Joe Wilson did. The bipartisan Senate Intelligence Committee investigation found that

The former ambassador was selected for the 1999 trip after his wife mentioned to her supervisors that her husband was planning a business trip to Niger in the near future and might be willing to use his contacts in the region"

and

An INR analyst's notes indicate that the meeting was "apparently convened by [the former ambassador's] wife who had the idea to dispatch [him] to use his contacts to sort out the Iraq-Niger uranium issue."

and (most important of all)

The report on the former ambassador's trip to Niger, disseminated in March 2002, did not change any analysts' assessments of the Iraq-Niger uranium deal. For most analysts, the information in the report lent more credibility to the original Central Intelligence Agency (CIA) reports on the uranium deal...

That is, contrary to what Wilson claimed, his wife was involved in arranging his trip and he did not uncover any information suggesting that Bush lied. Instead, Joe Wilson lied. Lots of left-leaning senators, including then vice-presidential candidate John Edwards, signed this report.

Why those on the left, who have many legitimate gripes about Bush's policies, would continue to champion Joseph Wilson's (long lost) cause has therefore long been a real mystery to me. It is all the more mysterious in light of the fact that it has also been strongly suspected for a very long time by anyone with more than a passing curiosity about the Valerie Plame story that Richard Armitage, not Karl Rove or Dick Cheney, leaked Plame's name to the press. Novak and Woodward, the initial recipients of the leak, have been dropping hints for a very long time that the leaker was not someone that the left would regard as someone with a political agenda. It therefore comes as something less than a complete shock to find this article today by Michael Isakoff:

Armitage, a well-known gossip who loves to dish and receive juicy tidbits about Washington characters, apparently hadn't thought through the possible implications of telling Novak about Plame's identity. "I'm afraid I may be the guy that caused this whole thing," he later told Carl Ford Jr., State's intelligence chief. Ford says Armitage admitted to him that he had "slipped up" and told Novak more than he should have.

and

As it turned out, Novak wasn't the only person Armitage talked to about Plame. Washington Post reporter Bob Woodward has also said he was told of Plame's identity in June 2003.

The only surprise is why so many on the left continued to cherish a nefarious scenario according to which Bush and his minions (Rove and/or Cheney) outed Valerie Plame because her husband had boldly revealed that Bush misled the nation when he spoke his famous 16 words: “The British Government has learned that Saddam Hussein recently sought significant quantities of uranium from Africa.” It has been obvious for a long time now that everything about the nefarious scenario -- so deeply believed by so many on the left -- was completely wrong. In fact, it has been clear for a long time that, in light of the facts, the nefarious scenario was much more indicative of the mentality and character of the true believers on the left than it was of Bush's mentality or character. Doggedly maintaining a belief that has been contradicted by factual information is indicative of something. But what?

Bush and many of his supporters may have been wrong about Saddam's WMDs. But the Senate Intelligence Committee, in the report I linked to above, examined the information available at the time and concluded that it was reasonable to believe that Saddam did have WMDs. Will advocates of the evil-Bush-vindictively-outed-Valerie-Plame scenario go back and examine whether the facts ever justified that story? If they do, they'll find that the facts never did. Thus, if you were one of those believers, you need to consider where your fact-free conviction came from. That conviction is indicative of something, and, no, it is not indicative of something about Bush. It is indicative of something about you. The Valerie Plame story is a psychological mirror for you to use, not final proof that Bush is the devious and spiteful man you always thought he was.

What you Need to Know about the Death Penalty

If you could save 18 innocent people from dying a violent death by simply flipping a switch, would you do so? Of course you would. But what if that switch activated an electric chair that strapped down a convicted killer? Would you save those innocent lives by ending a guilty one, or would you let them die and spare the killer's life?

If you are opposed to the death penalty, my guess is that you have always managed to avoid this moral dilemma based on what you know about the deterrent effect of the death penalty. What you think you know is that executions do not deter murder. What you need to know is that every recent study that has looked into this matter suggests otherwise.

Before your head explodes in a spasm of disbelief, you should know that I've done my homework on this issue. I cite the actual studies in question at the end of this post. I've read one of those studies in detail, and I have perused several others. And I also read a terrific review article entitled "THE DEATH PENALTY MEETS SOCIAL SCIENCE: Deterrence and Jury Behavior Under New Scrutiny" by Robert Weisberg of the Stanford University School of Law. His review was published in the Annual Review of Law and Social Science last year. In that review, he says this:

Recently, an impressive new generation of deterrence studies has promised to overcome these difficulties by relying on panel data—that is, data from numerous units (in terms of American criminal justice, the 50 states or all counties in the United States) for numerous time periods. These data sets allow for comparisons across jurisdictions over time; they typically include information on potentially confounding variables; they have enough observations to ensure that analyses based on them will have reasonable statistical power; and they benefit from the increased rate at which executions occurred during the 1980s and 1990s. And, most dramatically, these recent studies, using modern regression techniques, find that executions have not just a significant but a substantial deterrent effect.

I assume this is news to you. It certainly was to me.

The studies are extremely detailed, and they involve complex quantitative analyses. As such, skeptics will always be able to find fault with them. And they have (see here, for example). In light of recent critiques, Weisberg concludes that "...the relationship between executions and murders still lacks clear proof." The evidence might not yet amount to proof such that even skeptics would be persuaded, but it is no longer viable to assume, as many do, that social science has proven the opposite (namely, that the death penalty serves no deterrent effect). Instead, a whole pile of recent research points in the opposite direction, and skeptics are scrambling to fight them off. That seems like a tectonic shift to me.

There are many related issues that might determine one's stand on the death penalty (e.g., innocent people are sometimes executed, capital punishment may be applied disproportionately as a function of race, etc.), but these new findings should at least prompt death penalty opponents to say what they would do if faced with the moral dilemma that began my post. Notable scholars like Cass Sunstein are even asking Is Capital Punishment Morally Required? That question seems fair to me. Would you still oppose the death penalty even if each execution saved as many as 18 innocent lives, as some studies suggest is the case?

References:

Dezhbaksh H, Rubin P, Shepherd J. 2002. Does capital punishment have a deterrent effect? New evidence from post-moratorium panel data. Am. Law Econ. Rev. 5(2):344–76

Shepherd J. 2004a. Murder of passion, execution delays, and the deterrence of capital punishment. J. Leg. Stud. 33(2):283–322

Zimmerman P. 2004. State execution, deterrence, and the incidence of murder. J. Appl. Econ. 7:163–93

Cloninger DO, Marchesini R. 2001. Executions and deterrence: a quasi-controlled group experiment. Appl. Econ. 35(5):569–76

Brumm H, Cloninger D. 1996. Perceived risk of punishment and the commission of homicides: a covariance structural analysis. J. Econ. Behav. Org. 31:1–11

Ehrlich I, Liu Z. 1999. Sensitivity analysis of the deterrence hypothesis: let's keep the econ in econometrics. J. Law Econ. 41(1):455–88

Liu Z. 2004. Capital punishment and the deterrence hypothesis: some new insights and empirical evidence. East. Econ. J. 30(2):237–58

Yunker J. 2002. A new statistical analysis of capital punishment incorporating U.S. postmoratorium data. Soc. Sci. Q. 82:297–311

Iran: The only question

Democrats and Republicans will naturally differ in their approach to Iran, with Democrats being more likely to favor something like "constructive engagement" for a longer period of time than Republicans. Still, everyone agrees that "carrot-and-sticks" diplomacy should be tried first (and so it will be).

If that approach works, everyone will be happy, so the only real question is this: what approach would you favor if you became convinced that Iran's actual plan was to develop a nuclear bomb no matter what they verbally agree to do in response to UN demands and no matter what the cost might be in terms of economic sanctions?

What then?

That's the only question, and the answer to that question probably separates your average Democrat from your average Republican. If the Iranians have that plan in mind, I suspect that most Democrats would let the them have their nuclear bomb, whereas most Republicans would support military action to prevent that from happening. Moreover, I suspect that most Democratic politicians, if confronted with my question, would just say that they don't deal with hypotheticals (an answer that I would interpret to mean "let them have their bomb and then hope they don't use it").

Iran has a tried-and-true formula to follow, and it's the one I mentioned above: reluctantly agree to give up programs that could lead to the development of a nuclear weapon, promptly ignore that agreement, and then deny inspectors access to key sites (while proceeding full steam ahead with a nuclear bomb-making program). The worst case scenario from their point of view is that the world will get its act together enough to impose some feeble economic sanctions that will be easy enough to circumvent (as Saddam Hussein ably demonstrated).

What if that's Iran's plan? That is, what if they intend to develop a nuclear bomb no matter what the mix of "carrots and sticks" might be, so long as those sticks consist of nothing more threatening than economic sanctions?

What then?

It seems like a fair question to ask in light of the fact that it is transparently obvious (to me, at least) that this is their plan. Like it or not, there is a good chance that it will come down to a choice between letting Iran have their long-sought nuclear bomb or taking military action against them. It's a tough choice, and you should not cast your vote for any politician who refuses to take a stand on that issue.

Making News out of Random Error

The media needs a short course in the concept of "error variance" (also known as "random error"). Simply put, it's hard to know what numbers mean when they are aggregated over too few observations. In a post below, I used the baseball analogy to illustrate this point. In baseball, you need to average over a lot of games to determine who the best team is because individual games tell you so little. The outcome of an individual game is determined by many factors unrelated to which team is better (e.g., luck, a good player being a bit under the weather, a pitcher having an off day, etc.). All of those factors together constitute random error. Not appreciating the role of random error makes for a lot of news that is actually non-news. The best example might be opinion polls, where random fluctuations are reported as astonishing new developments, especially during a presidential campaign (I'll post the details on that some other time). Another example, which has my attention at the moment, concerns casualties in Iraq.

I posted earlier on the situation with respect to civilians, which has worsened since the February bombing of the Golden Mosque in Samarra. In the ensuing media frenzy, "civil war" was declared for the umteenth time, but the numbers actually show a stable level of increased violence (which is bad enough without repeatedly declaring an out-of-control spiral). Even in a stable situation, the numbers will fluctuate from month to month. That's the key point. But just as with opinion polls, where random fluctuations are reported as breaking news, random fluctuations in civilian casualties from month to month are severely overinterpreted. As you may have heard 100 times by now, July was the worst month ever with respect to civilian casualties in Iraq. True, but it was also much like the previous few months. June came in relatively low, July came in unusually high, and the best prediction for August (as for any future month) is that an average number of casualties will be observed.

Via Instapundit, I see that DailyDispatch is taking note of the fact that civilian casualties seem to be on the decline during the month of August, and this is being attributed to the new security crackdown in Baghdad. I am sure that the security crackdown is doing some good, but let's take a look at what would be expected during the month of August just based on the average of what we have seen since March:

The projected value for August (the blue bar) is simply the average of the red bars. That is, without looking at any numbers from August, the expected number of casualties for that month is 896. Well, August is three-quarters over, and we can look to the data supplied by Iraq Coalition Casualties to see what is happening so far. As of August 23, 607 civilian casualties have been documented. Projecting through August 31 based on what has happened so far yields a number of 818, which is not far from what one would expect. My point is that if it does turn out this way, then the month of August will be fairly typical. There is no "spiral of violence," and things in Iraq are not "escalating out of control." There is just a new and higher level of violence that was ushered in by the bombing of the Golden Mosque. Moreover, it is much too soon to attribute the improved August numbers to a security crackdown that is bearing fruit. The new security posture may be helping, and I sincerely hope that it is. But only time will tell, and not enough time has passed.

In the meantime, we can expect that random fluctuations will drive the news. In July, that worked against the American effort to democratize Iraq (because the news was so demoralizing). In August, if current trends hold, it will work in favor. But it all averages out in the end. To know what is really happening, you have to ignore the monthly random error and focus on the longer term.

Iraq: "Phased Redeployment" vs. "Stay the Course"

The major political decision facing the nation still has to do with Iraq and basically involves choosing between the "phased redeployment of US troops" vs. "stay the course." Ned Lamont epitomizes the former option at the moment (though Russ Feingold is probably the real standard bearer), whereas George Bush is a rock with respect to the latter.

It seems clear that violence against civilians in Iraq has increased substantially since the bombing of the Golden Mosque back in February, and many have expressed the concern that Iraq is on the verge of all out civil war. Moreover, virtually everyone agrees that 130,000 US troops are restraining what would be a vastly worse and prolonged spasm of violence were they to just pack up and leave. And this particular civil war, if you want to call it that, would not involve two armies facing each other down (as in the American Civil War). Instead, it would involve endless episodes of mass murder directed by various militias against large concentrations of civilians. That being the case, isn't it fair to describe the two political options we face as follows?

Option A ("phased withdrawal"): Let the unrestrained slaughter of innocent civilians in Iraq begin.

Option B ("stay the course"): Keep the troops in Iraq until they can facilitate the political climate and train the Iraqi security forces needed to keep that slaughter from ever happening.

It is not surprising to me that most people who favored the invasion of Iraq in 2003 favor option B now. What is surprising is how few people who opposed the invasion (often on humanitarian grounds) favor option B now. Having opposed the invasion on humanitarian grounds while favoring the "stay the course" option now (also on humanitarian grounds) seems like an entirely consistent -- perhaps even noble -- political position to me. As such, I'm puzzled by the near absence of people who hold that view.

Civilian Casualties in Iraq

A recent UN report showing that more than 3000 Iraqi civilians were killed in the month of July has cemented a widespread impression that violence in Iraq is "spiraling out of control." The UN report combined the number of civilians killed according to the Iraq Ministry of Health with the number of bodies reported by the Baghdad morgue. Many news stories have noted that July had the highest morgue tally ever (1885 bodies, often reported as "nearly 2000" for dramatic effect). Does that mean that things actually are spiraling out of control in Iraq, with each month vastly worse than the preceding one? No. In reality, civilian casualties have increased by about 50% since the bombing of the Golden Mosque in late February. The subsequent 5-month period of March through July might be referred to as the "civil war" period. While the level of violence against civilians clearly increased during that time, it has remained essentially stable over the last 3 months (which is hard to reconcile with an out-of-control spiral). My analysis of the situation is presented below.

I consider the Baghdad morgue figures in my analysis even though I am not sure that they are trustworthy in light of a July 25, 2006 article in NEWSWEEK, which had this to say:

The morgue is at the heart of that debate, because whoever controls the morgue controls the numbers. That person is radical Shiite cleric Moqtada al-Sadr. One day last week, a NEWSWEEK reporter saw more than a dozen militiamen, dressed in the traditional black of Sadr’s army, patrolling the facilities, keeping an eye on the staff. According to morgue employees, Sadr’s Mahdi militiamen aim to control the flow of information to give Sadr a leg up in the propaganda war.

Moqtada al-Sadr may be providing an undercount (to hide how many Sunnis he is really killing) or an overcount (to instantly generate demoralizing headlines in America like "Sectarian Carnage in Iraq"). Or perhaps he is just an honest reporter. But just to be on the safe side, I analyzed the numbers in two ways, one of which includes the morgue figures and one of which does not. The bottom line is that sectarian violence against civilians worsened after the bombing of the Golden Mosque in Samarra back in late February. But the evidence suggests that it increased to a new, stable level. It is not an out-of-control spiral. Neither scenario can be considered good news, but the difference between a stable level of violence that is 50% worse than it was and an out-of-control spiral of violence is important.

Several web sites are devoted to tracking the situation in Iraq, and one of those is Iraq Body Count (IBC). The numbers provided by IBC are based only on news reports and morgue statistics. They do not record all deaths, but their measures are sensitive to change (i.e., when things get worse, that shows up in their data). Their numbers document an undeniable increase in the level of violence against civilians over the last 5 months (i.e., since the bombing of the Golden Mosque). But those numbers also document the fact that the violence has not doubled or tripled and is not spiraling out of control. That is the key point I wish to make.

Here is what their numbers suggest in terms of the number of civilians killed, starting in March of 2005 (see technical footnote at the end of this post):

The most recent 5 months (shown in red) exhibits an increase over the two previous 5-month periods. The average number of deaths during the red period is 1665. The average during the green period was 1349 and the average during the blue period had improved to 1152. But the mosque bombing appears to have reversed that trend. The important point, though, is that the numbers are not spiraling out of control. It is more like a step function. That is, the mosque bombing ushered in a new and higher level of violence, but it did not unleash a spiral of violence. The last 3 months look fairly stable, and the information available thus far suggests that August will be lower July (reinforcing the notion that a stable level exists).

I also checked the numbers in the Iraq Coalition Casualties database, which, like the IBC, relies on media reports. However, they do not include the morgue figures (which is good because those figures may not be completely trustworthy), so their numbers are lower than those reported by IBC. Still, the same pattern is evident in their data. That is, there is a step-function increase beginning in March, but no subsequent spiral:

For both databases, July of 2006 was, indeed, the deadliest month, as many reports in the media have suggested. But -- and this is the key point -- it was not whoppingly different from May or June of this year. I expect that numbers for August will be right about at the average of the last 5 months (though one can never rule out a catastrophic increase in violence over the next 10 days or so).

Based on my previous post and this one, the bottom line appears to be the following:

The insurgency is showing signs of weakening (based on reduced casualties for US troops and Iraqi military/police), but there is a new, higher, and essentially stable level of civilian deaths following the bombing of the Golden Mosque in late February. Remember that the next time you read that the insurgency is gaining strength and that violence is "spiraling out of control" in Iraq.

Footnote: For both analyses, I removed 1000 casualties from the August-2005 figures because they were caused by a stampede. Also, for the IBC analysis, I added in the Baghdad morgue figures for May, June and July of this year because they have not yet been added to the IBC database. The morgue numbers for those months have been reported by the BBC, so I just added the same percentage of morgue casualties that were added in prior months (they don't add all of them because some of them have already been added to the database based on news reports, and they try to avoid redundancy). Also, 923 casualties were reported by the IBC to have occurred between early April and the beginning of August of 2005, but they could not be assigned to particular months. I divided that total by 4, which comes to 231 per month, and added that number to the May, June and July 2005 numbers from their database.

The Strength of the Insurgency in Iraq

A New York Times article from a couple of days ago erred badly. I'm surprised no one has called them on it. They reported that attacks on American and Iraqi forces have doubled and claimed that the insurgency is clearly on the increase:

WASHINGTON, Aug. 16 — The number of roadside bombs planted in Iraq rose in July to the highest monthly total of the war, offering more evidence that the anti-American insurgency has continued to strengthen despite the killing of the terrorist leader Abu Musab al-Zarqawi.

and:

“The insurgency has gotten worse by almost all measures, with insurgent attacks at historically high levels,” said a senior Defense Department official who agreed to discuss the issue only on condition of anonymity because he was not authorized to speak for attribution.

The idea that the insurgency has gotten worse "by almost all measures" is amazingly off the mark. In fact, it is not worse according to the most obvious measures. To appreciate why this is so, you need to understand that measures taken over the short term are inherently noisy. So, looking at one month of casualty figures tells you very little in the same way that looking at the outcome of a single baseball game tells you very little about the strength of the two teams. To get a clear picture, you need to aggregate the numbers over time. Since we have complete data for the first 7 months of 2006 (January through July), I looked at what has happened over the first 7 months of this year compared to the immediately preceding 7 months (i.e., the last 7 months of 2005). Here is what I found:

US military fatalities are down in the first 7 months of 2006 (average of 57 per month) compared to last 7 months of 2005 (average of 73 per month). This is easily confirmed by going
here.

In addition, the number of US fatalities from IEDs is down in the first 7 months of 2006 (average of 32 per month) compared to last 7 months of 2005 (average of 42 per month), this despite the fact that more IEDs are being placed (incompetently, perhaps). This is easily confirmed by going here.

In addition, Iraqi military and police casualties are down in the first 7 months of 2006 (average of 177 per month) compared to last 7 months of 2005 (average of 243 per month). Again, you can confirm this yourself by going here.

It is hard to argue that the insurgency is gaining in strength, as the New York Times does, unless you ignore the most obvious measures (while ironically claiming that things are getting worse by almost all measures).

The NY Times article goes on to say:

The increased attacks have taken their toll. While the number of Americans killed in action per month has declined slightly — to 38 killed in action in July, from 42 in January, in part reflecting improvements in armor and other defenses — the number of Americans wounded has soared, to 518 in July from 287 in January. Explosive devices accounted for slightly more than half the deaths.

Is the number of wounded really soaring? In truth, the number of US wounded is down in the first 7 months of 2006 (average of 425 per month) compared to last 7 months of 2005 (average of 499 per month). This is easily confirmed by going here.

I am surprised that the New York Times cannot even be trusted to consult readily available data that can be easily examined to evaluate their judgment of the strength of the insurgency. At the very least, they should have a paragraph in their article that begins like this: "Despite many figures suggesting that the strength of the insurgency is, if anything, declining, our belief is that it is actually gaining in strength because the number of IEDs being planted is on the increase. The reason why many other indicators suggest a weakening insurgency is that..."

I have no idea how they'd complete that sentence, but that's what the reporters need to do if they wish to grapple with the actual evidence (big "if" there).

Later, I'll discuss what I've learned about civilian casualties in Iraq by performing similar analyses.

First try

What the world least needs right now is another blogger. And with that opinion, so begins my blogging career.

Engram