October in Iraq

Every month I try to bring you an accurate and non-hysterical overview of the violence in Iraq. You already know that October was a tough month, but I compiled casualty statistics from Iraq Casualty Count (ICC) to get some idea of how bad it was compared to earlier months. ICC tracks civilian deaths reported by the media. This method that does not detect all deaths (they probably catch 1/3 to 1/2 of them) but it does provide a useful metric that is sensitive to change. As such, when things get better, it shows up in the casualty counts, and the same holds true when things get worse (as they did in October). I don't believe there is a better way to follow the trends in Iraq.

As I am sure you know, the February 2006 bombing of the holy Shiite mosque in Samarra marked a turning point in Iraq. By that act, al Qaeda made good on its promise to provoke the Shiite militias into sectarian conflict with the Sunnis. Here are the monthly civilian casualty counts, beginning 4 months before the bombing (the blue bars) and continuing for the next 8 months:


As you can see in the figure, the first 4 months after the bombing (shown in orange) reflect a noticeable increase in casualties, and the next 4 months (shown in red) reflect a further increase. [note: if you go to the ICC web site, you'll see that the raw numbers in the relevant table appear to show a big drop in casualties for October, but that is an illusion caused by their odd decision to include some extra casualty counts from the morgue in August and September only -- those reports need to be subtracted out so that the same method is used for all months]. Here are the casualty counts averaged over 4 month blocks to make the trend easier to see:


Will the violence keep increasing? That is, are things finally spiraling out of control? No one knows, but my guess is that Prime Minister Maliki has decided that the violence inflicted on Sunnis by Muqtada al Sadr's Mahdi Army is serving a useful purpose. Al Qaeda thinks so, too, which is why they keep bombing markets and mosques to further provoke the Shiites (at least I think it is al Qaeda doing that, not the Sunni insurgents as I previously thought). Both al Qaeda and Maliki want to see Sunnis under assault, but they have placed different bets on how the Sunnis will respond. Al Qaeda thinks that the Sunnis will further align themselves with that terrorist organization in order to defend themselves against the Shiites more effectively. Maliki thinks that the Sunnis will instead appeal to the government for help and perhaps agree to give up the insurgency in return for the Shiite militias being disarmed and purged from the police force. A lot more violence against Sunni civilians will probably have to occur before the Sunnis make up their minds one way or the other, so I don't expect a quick drop in these casualty figures.

While al Qaeda and Prime Minister Maliki use Sunnis as pawns, the Sunnis themselves are focused on waging an insurgency against American and Iraqi troops. On that front, things seem mostly stable. For example, Iraqi police and military units have not seen their casualties increase very much in recent months, though they increased somewhat in the most recent 4-month period:


A similar story holds for US military casualties. This month was a particularly bad one (with 103 killed), but that probably reflects a mini Tet offensive designed to induce Americans to vote for Democrats (not a sudden strengthening of the insurgency). Averaged over 4-month blocks, we see no particular trend:


Earlier I had argued that the casualty statistics suggest a slight weakening of the insurgency, but its strength seems fairly constant judging by these measures. I imagine that this will continue for some time to come, and the inability of American and Iraqi forces to put the insurgency down probably explains why Prime Minister Maliki has decided to allow Muqtada al Sadr's Mahdi Army to eradicate Sunni males by the dozens every few days. It's a way to finally bring the Sunnis to the negotiating table, though no one can say that publicly because it's too barbaric for civilized folks to openly acknowledge.

One Week to Go

With one week to go before the election, I detect no change at all in the predicted gain of 5 plus-or-minus 1 Senate seats for the Democrats. A few of the individual races have shifted around a bit, but the overall story remains much the same. Here is the latest snapshot based on averaged polls (the only kind of polls worth attending to) and the betting at Tradesports.com:


The states shown in red are currently held by Republicans and those shown in blue are currently held by Democrats. As has long been true, the Democrats are on track to hold on to New Jersey and to pick up 4 more seats (Montana, Ohio, Pennsylvania, and Rhode Island). Missouri remains a toss-up, whereas Tennessee now looks likely to remain in Republican hands. The betting on the individual Senate races at Tradesports.com (also shown in the table above) tells pretty much the same story.

The surprise is that Virginia is now in the toss-up category. I guess voters did not like George Allen's strategy of quoting explicit passages from Webb's novels.

There is a lot of talk about polls being off the mark this year, either because conservatives are disproportionately refusing to talk to pollsters or because cell phone users are not be sampled or because Democrats are being oversampled for one reason or another. I hope some of that is true, but I don't think it is. Unless something dramatic happens this week, what has seemed true for a long time still seems true: the Democrats are going to pick up at least 4 seats in the Senate and they'll probably take control of the House. At Tradesports.com, the current betting is that the Republicans have a 67% chance of holding on to the Senate but only a 30% chance of holding on to the House. Seems about right to me.

At War with Al Qaeda in Iraq

The recent upheaval in Iraq is commonly referred to as sectarian violence, but it seems that a more accurate way to put it is that al Qaeda is attacking Shiite civilians so that Shiite militias will respond by attacking Sunni civilians.

This story from Fox News characterizes the situation in a way that seems accurate to me:

The explosion in the sprawling Shiite slum of Sadr City tore through food stalls and kiosks at 6:15 a.m. (0315 GMT), cutting down men who gather there each morning hoping to be hired as construction workers. At least 59 people were wounded, police Maj. Hashim al-Yasiri said.

Sadr City, is a stronghold of the Mahdi Army loyal to radical Shiite cleric Muqtada al-Sadr, and has been the scene of repeated bomb attacks by suspected al-Qaida fighters seeking to incite Shiite revenge attacks and drag the country into full-blown civil war.

Why would al Qaeda want a full-blown civil war? The article does not say, but Zarqawi's 2004 letter does. Al Qaeda launches terror attacks against Shiite civilians because they want the Shiite militias to retaliate against Sunni civilians which, in turn, will push the Sunnis to further ally themselves with al Qaeda. Once that happens, al Qaeda will have more bases of operation in Iraq (i.e., more places to hide), which they need to wage their holy war. They want war with America and with the Shiites, and Iraq is the place to do it. But they need places to hide, and only the Sunnis can provide that for them. This analysis fits with a story from today's Washington Times (via Powerline):

Saleh Mutlaq, whose Iraqi National Dialogue group supports the U.S.-backed political process, said al Qaeda's growing control of strongholds at the heart of the Sunni insurgency was paving the way for an Islamic fundamentalist state in western and central Iraq, and even Baghdad.
...
He said al Qaeda plans to expand its territory both in Iraq and beyond its borders. "They want to capture territory to attract more jihadists to destabilize everywhere, and they think they can take over all of Iraq later when they have the territory to operate," he said.

Al Qaeda's vision for Iraq needs to be kept in mind as we contemplate empowering Democrats in next week's elections. Iraq is obviously the central front in the war on terror, yet the Democrats mainly want to withdraw from it. If the Democrats do regain power next week, I hope they will think twice about working to hand a glorious victory to al Qaeda by insisting that we set a timetable for withdrawal from Iraq.

The Democrats Continue to Amaze

A CNN news story is called "Democrats: To fix Iraq, first GOP must go."

RICHMOND, Virginia (AP) -- The only remedy to a series of Iraq policy failures by President Bush is a Democratic takeover of Congress in the November 7 election, Virginia Senate candidate Jim Webb said Saturday.
...
"A Democratic Congress will demand from day one that the president find a real way forward in Iraq. We'll work with the administration and other Republicans to develop a concrete plan, but none of us are ready to settle for empty rhetoric, or the same old unacceptable results," Webb said.

I'm looking everywhere for a shred of an idea from the Democrats, and it just boggles my mind that the only plan that they have is to stand on the sidelines and carp at what the Republicans are doing. That is not entirely useless, but their ability to do that is not a reason to turn the House or the Senate over to them.

To say that Bush is incompetent in the middle of this war and that the solution is to get rid of the Republicans is like calling a neurosurgeon incompetent in the middle of an operation and suggesting that the solution is to get rid of the doctor. Well, maybe that is the solution, but wouldn't you want to know what the plan is before hauling the surgeon off to jail? I mean, the patient is lying there on the table with his head cut open, and something needs to be done. Obviously, you need to keep the surgeon in place until you can figure out how to finish the operation without him. Same goes for Iraq.

So how are the Democrats going to finish that operation? Like this, apparently: "A Democratic Congress will demand from day one that the president find a real way forward in Iraq."

That's inspiring. Makes me want to rush right out and give the Democrats the power to cut off funding for Iraq and to impeach Bush so we can implement the Kerry/Murtha plan to walk away from Iraq in the middle of an operation.

Michael Barone Weighs in on the Polls

The blogosphere is abuzz with a new column by Michael Barone in which he notes that current polls seem to oversample Democrats, so things may be brighter for the Republicans than they seem. Instapundit, Powerline and the NRO Corner all link to it.

My only problem with this is that it is a lot like all the talk we heard about before the last presidential election, yet the polls turned out to be dead-on accurate. That fact seems relevant to the current discussion.

Here is some pre-election poll analysis from a pro-Kerry columnist explaining why the polls were wrong to predict that Bush would win the 2004 election:

Why Kerry will beat Bush
November 2, 2004
...
Finally, one thing kept clawing at you. Cell phones. Long I have wondered how many there were. Everybody I know, smart people, politicians, news directors, thought that there were, oh, 40 million or so. I call the cell phone institute in Washington last Sept. 12. They told me that there were 165 million cell phones in use in the United States, That is 165,000,000. One month later, I asked again. It was up to 170 million -- 170,000,000. Yes, a great number also had land lines. But of this 170 million cell phone users there were 40 million between the ages of 18 and 29, and these people usually have no other phones.
...
Not one cell phone in the United States had been reached by a political poll. These old-line poll takers don't know who cell phone users were or where they lived.

So you were getting CBS/New York Times polls proclaimed as most important and real. One hundred seventy million cell phones and you don't poll one of them. The polls they are pushing at you in the news magazines, on the networks, in the big papers, are such cheap, meaningless blatant lies, that some of these television stations should have their licenses challenged.
...
Then a night or so ago, somebody finally tried a poll of cell phone users between the ages of 18 and 29. John Zogby conducted the survey in conjunction with Rock the Vote and the results showed Kerry at 55 percent and Bush at 40.

Then the Kerry people ran their own poll, which took a lot of work. It was the first time they had reached any cell phone users. The result was Kerry 59 and Bush 39.

Remember that? There was endless talk about missing cell phone users, which made me really curious to find out if pollsters were too dumb to appreciate the possible effects of that. And here is LGF suggesting that a Newsweek poll before the last presidential election oversampled Democrats:

Did Newsweek choose a lower percentage of Republicans for the first debate to set up Kerry’s “comeback,” or did they stack the deck with more Democrats in the second poll?

There was endless discussion about this, too, which is why I waited with bated breath not only to find out who would win the election but also to see whether the averaged polls would turn out to be accurate. In the end, they were, which means that the apparent disagreements between various polls and the differences in the number of Democrats and Republicans sampled was just due to unavoidable random error. The temptation to explain away polls that you don't like is powerful, and it affects those on the left as well as the right. But polls have a track record that is too good for me to dismiss what they have to say. Polls that survey all adults are biased (because they include many people who do not vote, and most of those people lean left), but polls of likely voters have to be taken seriously.

It could be that things are different this time around and that now the pollsters really are oversampling Democrats, but I don't believe it. Not yet anyway. If the Republicans do better than predicted (as I hope they will because my own party is clearly not ready for prime time), I suspect that it will show up in the polls and in the betting before November 7. That is, they will do better because of a real change in how people are inclined to vote, not because pollsters are systematically biased.

Then again, this is one of the reasons why I am so intrigued by the upcoming election. It is the second time that I'll be able to test the validity of averaged polls (of likely voters), and I can hardly wait to see how it turns out. It will be easiest to tell for the Senate races because we have more polling information for them than for the House races. For the moment, though, I remain of the opinion that all of the pre-election poll analyses are off the mark, whereas the polls themselves are on the mark. I'll update that opinion on November 8.

Jack Kelly: We Must Get Rid of al-Sadr

In a new column, Jack Kelly becomes the latest to argue that we need to terminate Muqtada al-Sadr, and he repeatedly asks "Why is Moqtada al-Sadr still alive?"

Anyone who asks that question should at least consider this possible answer:

Because he is bad for the Sunni insurgency.

He is, in fact, their biggest threat, by far. His Mahdi Army is the reason why the Sunnis are much less interested in seeing American troops leave their country than they were before. Eliminating al Sadr and disarming his militia is the biggest gift we could possibly give to the Sunni insurgency that is fighting to undo everything we have done in Iraq.

That's why he's still alive.

What the Dems Would Do

I've been following discussion in the media about what the Democrats would do if they regain power on November 7. In this post, I consider 4 recent articles on that topic (3 of which I've considered in prior posts). Most of the discussion is by writers who are sympathetic to the left wing agenda, and even they find the Democratic plan to be dyspeptic at best. Shouldn't someone be impressed by what the Democrats have to offer before we put them in charge of anything?

Consider again this article by Michael Isikoff and Holly Bailey (with Mark Hosenball and Eleanor Clift) from the October 30 issue of Newsweek:

What the Dems Would Do
...
Democratic leaders are largely sticking with domestic issues in part because they have yet to come up with a coherent plan for the biggest problem of all: articulating a clear way out of Iraq. On the campaign trail, Democrats have been content to bash Republican failures and say they'd do better. The Democrats' official line is to promise a "new direction," and to urge, vaguely, "redeployment." If they win, they'll be forced to say what, if anything, that means.

What does "redeployment" really mean? Nothing, as we see in this sympathetic column by David Broder from last week:

If the Democrats take Congress, here's what they'd do

No one speaks more authoritatively for the Democrats on defense and national-security issues than Sens. Carl Levin of Michigan and Jack Reed of Rhode Island, both longtime members of the Armed Services Committee.
...
On Iraq, the two Democrats harked back to the amendment that 39 senators supported during a debate earlier this year — an amendment that called for a start on U.S. troop withdrawals within six months, but set no numbers and specified no target date for ending the U.S. military presence.

Reed, who has made many trips to Iraq and returned just weeks ago from his most recent visit, described the "very, very difficult situation" he found there. "We have to begin to work toward redeployment without setting a timetable," he said. "We have to start laying out some red lines for the Iraqis ... give them some clear goals we want them to achieve."

That really clears up the Democratic plan for Iraq, doesn't it? I mean, who could have any trouble understanding what a Democrat means when he says that we should redeploy without setting a timetable? Obviously, the Democrats want to abandon Iraq, they know that Americans are frustrated with the continuing violence in Iraq, but they read polls suggesting that people are not wild about setting timetables for withdrawal. So they've come up with a "plan" that does little more than summarize these feelings. An editorial in today's Washington Post -- a relatively sympathetic outlet for the Democrats -- again highlights the obvious problem with these vacuous ideas:

What Democrats Would Do
Their agenda is a hodgepodge of good ideas, bad ideas and no ideas.
...
But with it looking increasingly as if Democrats, after 12 years in the minority, will take over the House at least, it's worth looking at their stated agenda -- "A New Direction for America" -- for a glimpse at what a Democratic majority might entail.
...
On national security, the House Democrats' plan offers more goals than details. Who could disagree with promises to "eliminate Osama Bin Laden, destroy terrorist networks like al-Qaeda, finish the job in Afghanistan and end the threat posed by the Taliban" or "redouble efforts to stop nuclear weapons development in Iran and North Korea?" But the hard part -- on which Democrats offer no details -- is how that is to be done.

On Iraq in particular, the agenda calls for "the responsible redeployment of U.S. forces," with "Iraqis assuming primary responsibility for securing and governing their country." Again, what's missing are the details of what "responsible redeployment" might look like. "Insist that Iraqis make the political compromises necessary to unite their country and defeat the insurgency," the Democrats say. Okay, what if that insistence doesn't yield the desired result?

As I have said before, walking into a Democratic caucus would be walking into an angry, idea-free zone. But it looks like we are going to put them in power anyway. The editors go on to say:

The agenda is heavy on ideas -- raising the minimum wage, letting the government negotiate Medicare drug prices -- that may have more popular appeal than real-world impact, though we agree it's past time for the minimum wage to go up. Other proposals sound good but are bad policy. Screening 100 percent of cargo containers at the point of origin would be expensive, time-consuming and impractical. The vow to "prevent outsourcing of critical components of our national security infrastructure -- such as ports, airports and mass transit -- to foreign interests that put America at risk" is more Dubai ports demagoguery.

If a bomb exploded in a port tomorrow (which I sincerely hope does not happen), it would ensure a Republican victory on November 7 despite all this talk by Democrats about enhancing port security. As you can see from all of these editorials, no one takes the Democrats seriously -- not even a largely sympathetic media. Terrorists who might be thinking of launching an attack on one of our ports are not going to take them seriously either.

America is not poised to vote for Democrats because they like their ideas. As I explained here, it is just the 6th-year itch. Americans are ready for a change and, unfortunately, my own party is the only alternative they have. As such, it looks like the Dems will get a chance to make the moves that Isakoff et al. describe in their Newsweek article:

If Democrats take it back next month, the party will once again be in charge of all the committees. Dingell—now 80 years old and more ornery than ever—is all but certain to return to his old job [chairman of the House Energy and Commerce Committee]. After 12 long, frustrating years as the panel's ranking minority member, a title that left him little more than the power to complain, nothing animates Dingell more than the thought of making up for lost time.
...
Henry Waxman of California is another Democratic old-timer whose ire never dims. A tireless investigator, he's in line to head the Government Reform Committee, and plans to take aim at Halliburton and alleged rip-offs and contract abuse in Iraq. Then there's Charles Rangel, the New York congressman who's never met a cable show he didn't like. He is set to take over the Ways and Means Committee, and wants to take a hard look at the Bush tax cuts. John Conyers of Michigan has waited for years to head the Judiciary Committee. He's likely to convene hearings on the Patriot Act and domestic wiretapping. In the past, he has suggested the possibility of impeachment hearings for President Bush.

Pete DuPont has yet another article on this topic as well in which he asks Republicans deserve to lose, but what happens if Democrats win? Here are his answers:

First, Minority Leader Nancy Pelosi has promised that election of a Democratic House would insure "a rollback of the [Bush] tax cuts." Rep. Charles Rangel of New York, who would be chairman of the Ways and Means Committee, would make sure no tax cut extension bill would ever get to the floor. He voted against the 2001 and 2003 Bush tax cuts and the bill that later extended the tax cuts until 2010 (as did all but seven of the 205 Democratic House members). In September Mr. Rangel said that he "cannot think of one" Bush tax cut he would agree to renew.
...
As for the war in Iraq, Mr. Rangel observed that "You've got to be able to pay for the war, don't you?" In other words, end it by simply defunding it. Rep. John Murtha of Pennsylvania calls for "immediate redeployment of U.S. troops" and intends to run for majority leader if the Democrats take control of the House. Ninety percent of House Democrats opposed the terrorist surveillance program, and 80% voted against the recent terrorist interrogation legislation.

Finally, when we see what the new leaders of a Democratic House are likely to do, their views are--well--very different from most Americans. Rep Henry Waxman of California would become the Government Reform Committee chairman, and believes domestic terrorist surveillance is "illegal." He would use his subpoena power to launch investigations to try and limit the president's anti-terrorism powers.

Rep. John Conyers of Michigan, who would become chairman of the Judiciary Committee, has talked about subpoenaing "Bush administration officials to answer questions and face the consequences for their abuses of power." In other words, impeachment.

Speaker-to-be Nancy Pelosi has indicated she would like to put Rep. Alcee Hastings of Florida in charge of the House Intelligence Committee. As a federal judge, he was impeached in the House by a 413-3 vote, and removed from the bench by the Senate for bribery, corruption, and perjury. Rep. Hastings would lead the oversight of America's anti-terrorism policies.

It's ugly. It's real ugly. Even so, America seems poised to turn the House of Representatives over to my once serious and substantive party. Again I say:

Good heavens, America. What are you thinking?

Revisiting the 2004 Zarqawi Letter

As you may recall, coalition officials came across a letter written by Abu Musab al-Zarqawi back in early 2004. A translation of that letter appears here, and I just read it again in light of what has been happening in Iraq lately.

As everyone knows, al Qaeda bombed a holy mosque back in February of this year, and the Shiites have been retaliating against the Sunnis ever since. From Zarqawi's letter, it is clear that this was the plan. But why, exactly?

First, consider the central issue that Zarqawi is concerned about:

There is no doubt that our field of movement is shrinking and the grip around the throat of the Mujahidin has begun to tighten. With the spread of the army and police, our future is becoming frightening.
...
As we get closer to the decisive moment, we feel that our entity is spreading within the security void existing in Iraq, something that will allow us to secure bases on the ground, these bases that will be the jump start of a serious revival, god willing.

The basic issue is that al Qaeda's field of movement is shrinking as the Iraqi army and police become better trained, and al Qaeda needs more bases on the ground from which to wage jihad. What's the plan to obtain those bases? It's this:

The Shi'a in our opinion, these are the key to change. Targeting and striking their religious, political, and military symbols, will make them show their rage against the Sunnis and bear their inner vengeance. If we succeed in dragging them into a sectarian war, this will awaken the sleepy Sunnis who are fearful of destruction and death at the hands of these Sabeans
...
So i say again, the only solution is to strike the religious, military, and other cadres of the Shi'a so that they revolt against the Sunnis.
...
If we are able to deal them blow after painful blow so that they engage in a battle, we will be able to reshuffle the cards so there will remain no value or influence for the ruling council, or even for the Americans who will enter into a second battle with the Shi'a. This is what we want. Then, the Sunni will have no choice but to support us in many of the Sunni regions.

That's the plan to secure more bases of operation for al Qaeda in Iraq. If the Sunnis feel threatened by the Shiites, they will, theoretically, turn more fully toward al Qaeda.

Zarqawi managed to incite the planned sectarian conflict with unbelievable effectiveness. Even the timing of the critical attack on the holy mosque was pre-planned:

The zero-hour needs to be at least four months before the new government gets in place. As we see we are racing time, and if we succeed, which we are hoping, we will turn the tables on them and thwart their plan.

Well, they bombed the mosque in February, which was about 3 and a half months before the government was finally formed. Ironically, this article says that the unity government was established on the very day that Zarqawi was killed:

The last step to completing the government came on June 8, the same day a U.S. air strike killed al-Qaida in Iraq leader Abu Musab al-Zarqawi, when al-Maliki filled the remaining three cabinet posts for defense, interior and national security.

Zarqawi was killed, but his plan -- as detailed in that captured letter from 2 years ago -- unfolded according to his vision. The mosque was bombed a few months before the government was in place, and Shiite militias are now showing their rage against the Sunnis.

Now that the Shiites have done their part, Zarqawi's letter predicts that the Sunnis will do theirs (i.e., "...the Sunni will have no choice but to support us in many of the Sunni regions"). But will the Sunnis react that way? That's the question that I believe is being resolved in Iraq right now.

Prime Minister Maliki did not ask for the sectarian war incited by Abu Musab al-Zarqawi, but my guess is that he is trying to use it to his advantage. The Shiite militias are on the war path, and Maliki's implicit message seems to be that they are going to stay that way unless the Sunnis reject al Qaeda and come to terms with the government.

The Sunnis have no good options. They can embrace al Qaeda, as Zarqawi envisioned, in which case they face the prospect of genocide at the hands of well organized and ruthlessly efficient Shiite militias, or they can come to terms with the unity government, in which case they will have to accept a deal that is much worse than what they had under Saddam. While they think it over, young Sunni males are being executed by the dozens every few days in Baghdad. Let's hope they come to their senses sooner rather than later.

Charlie Cook on the Maryland Senate Race

Paul has a new post at Powerline entitled "A Toss-Up in Maryland?" It's about how Charlie Cook has changed his call on the Maryland Senate race between Michael Steele and Ben Cardin from "leans Democratic" to "toss-up." I am surprised that a veteran poll watcher (i.e., Charlie Cook) says this:

Polling taken since the end of September has been inconsistent. Voter/Consumer Research, a Republican firm, has released two surveys. The first conducted September 27-28 of 602 likely voters, showed Cardin ahead by five points, 44 percent to 39 percent. The second taken this week (October 22-23 of 602 likely voters) had a narrower gap of just two points, 41 percent for Cardin and 39 percent for Steele. Another Republican poll, this one conducted by Public Opinion Strategies (October 2-4 of 800 likely voters) had Cardin leading, 47 percent to 43 percent.

Democrats are touting two surveys that show wider margins and have Cardin above 50 percent. The first is a USA Today /Gallup poll (September 27-October 1 of 678 likely voters) that gave Cardin a 15-point advantage, 54 percent to 39 percent. A Garin-Hart-Yang Research poll for the Democratic Senatorial Campaign Committee (October 23-24 of 602 likely voters) and released after Republicans put out their numbers had Cardin up by 12 points, 52 percent to 40 percent.

These differing results aren't easily explained so it's time to pull out the tried and true, patented Cook Political Report Law of Polling Averages. These five polls combine to give Cardin an average lead of 7.6 points, which would seem to make sense.

He seems to appreciate the value of averaging polls, but he does not seem to appreciate the fact that the apparently inconsistent story told by various polls -- with leads for Cardin ranging from as small as 2 points to as large as 15 points -- is exactly in line with what you'd expect if the true lead were about 7.6%. Charlie Cook and other political analysts should read my explanations here and here!

Michael Steele may very well have closed the gap (a recent ad by his sister, for example, is amazing), but the polls don't yet allow us to make that claim. Perhaps they will as time goes on, but, for now, you have to conclude that Cardin has a lead of 7 or 8 points. Charlie Cook's political instincts tell him that the race is closer than that, and I defer to him on that score (i.e., I do not trust my political instincts). I do have faith in averaged polls, though, which is why Maryland is not yet a toss-up in my book!

NPR's House Prediction

Someone at NPR has analyzed the House races and came up with this projection:


If we distribute the 20 toss-ups evenly, the prediction is that the Democrats will end up with 221 seats to 214 for Republicans. That's pretty close the average prediction of other independent analysts (which came to 224 for Democrats and 211 for the Republicans).

On the RealClearPolitics blog, John McIntyre says that there are two very different ways to look at these projections. One is that the Democratic are likely to take over the House in light of the fact that they are so far ahead in the polls. The other is that the polls are wrong because they over-sample Democrats, in which case it would be reasonable to assume that things are better for the Republicans than they seem.

I don't really buy the oversampling argument because, if it were true, then it would have been true during the last presidential election. Back then, though, averaged polls were dead-on accurate. I can't imagine why that wouldn't be true now.

If the Republicans manage to win both houses of congress, I expect to see evidence that they are going to win show up in the polls and in the betting as we move closer to election day. In the Senate, I see some evidence that things have shifted in direction of the Republicans. For example, the betting at Tradesports.com now shows a 73% chance of the Republicans hanging on (it was at 65% 10 days ago). So far, though, the House still looks likely to change hands (with only a 36% chance of the Republicans hanging on, which is nearly identical to the 35% value in effect 10 days ago).

al Qaeda's Media War and the CNN Sniper Video

Here's a relevant story from Reuter's that I found via Glenn Reynolds and Austin Bay:

U.S. foes ramp up media campaign in "war of ideas"

WASHINGTON (Reuters) - As U.S. military losses mount steadily in Iraq, a document issued by a group linked to al Qaeda spells out new goals for America's most determined enemies and calls for a media war against the United States.

The document, which began circulating on the Internet this month, illustrates the techniques Washington's enemy is using in what President George W. Bush has called the "war of ideas."

"The people of jihad need to carry out a media war parallel to the military war ... because we can observe the effect that the media have on nations," said the document, signed by Najd al-Rawi of the Global Islamic Media Front, a group associated with al Qaeda.

The author suggests that video of attacks on U.S. troops in Iraq could be a weapon in the media war and sway U.S. public opinion.

It's that last sentence that I thought would have more people talking. Al Qaeda thinks that showing a video of attacks on U.S. troops in Iraq could be a propaganda weapon in the media war and sway U.S. public opinion, yet CNN thinks it was right to show that insurgent-supplied sniper video a few days ago. As they put it:

CNN issued a statement saying that the decision was "a difficult one, but for a news organization, the right one. Our responsibility is to report the news."

I wonder if they still feel that way after the Reuters report that appears to show that they were played like fools? Well, from this testy exchange between Lynne Cheney and Wolf Blitzer tonight, it looks like CNN still has much to learn:

CHENEY: Right, but what is CNN doing? Running terrorist tape of terrorists shooting Americans. I mean, I thought [Rep.] Duncan Hunter asked you a very good question, and you didn't answer it. Do you want us to win?

WOLF: The answer of course is we want the United States to win. We are Americans. There's no doubt about that.

CHENEY: Then why are you running terrorist propaganda?

WOLF: Well all do respect, this is not terrorist propaganda.

CHENEY: Oh, wolf.

WOLF: This is reporting the news. Which is what we do, we are not partisan.

CHENEY: Where did you get the film?

WOLF: We got the film, look, this is an issue that has been widely discussed. This is an issue we reported on extensively. We make no apologies for showing that. That was a very carefully-considered decision why we did that. And I think, I think --

CHENEY: Well I think it's shocking.

WOLF: If you are a serious journalist, you want to report the news. Sometimes the news is good, sometimes the news isn't so good.

CHENEY: But wolf, there's a difference between news and terrorist propaganda.

Indeed there is. Wolf Blitzer boldly claims that the CNN sniper video is not terrorist propaganda, but the terrorists who supplied the tape beg to differ. I hope CNN does a story on this new al Qaeda document because I would love to hear what they have to say for themselves in light of it.

NPR's House Prediction

Someone at NPR has analyzed the House races and came up with this projection:


If we distribute the 20 toss-ups evenly, the prediction is that the Democrats will end up with 221 seats to 214 for Republicans. That's pretty close the average prediction of other independent analysts (which came to 224 for Democrats and 211 for the Republicans).

On the RealClearPolitics blog, John McIntyre says that there are two very different ways to look at these projections. One is that the Democratic are likely to take over the House in light of the fact that they are so far ahead in the polls. The other is that the polls are wrong because they over-sample Democrats, in which case it would be reasonable to assume that things are better for the Republicans than they seem.

I don't really buy the oversampling argument because, if it were true, then it would have been true during the last presidential election. Back then, though, averaged polls were dead-on accurate. I can't imagine why that wouldn't be true now.

If the Republicans manage to win both houses of congress, I expect to see evidence that they are going to win show up in the polls and in the betting as we move closer to election day. In the Senate, I see some evidence that things have shifted in direction of the Republicans. For example, the betting at Tradesports.com now shows a 73% chance of the Republicans hanging on (it was at 65% 10 days ago). So far, though, the House still looks likely to change hands (with only a 36% chance of the Republicans hanging on, which is nearly identical to the 35% value in effect 10 days ago).

Karl Rove and the House Races

I've been following the polls on the hotly contested Senate races fairly closely, and those polls lead me to predict a gain of 5 plus-or-minus-1 seats for the Democrats on November 7 (they need 6 to re-take the Senate). I haven't been following the House races as closely because there are just too many of them to keep track of. However, you can find more than you'd ever want to know about these races in a Wikipedia article here. The article contains quite a few numerical predictions made by various non-partisan groups (e.g., Larry Sabato's Crystal Ball, Cook Political Report, etc.) that are a bit too detailed for my tastes, so I boiled down the predictions into a single, easy-to-read table. I assume these analysts are basing their predictions on polls and financial records and such, but I don't really know. Here is what they predict:


As you can see, every single one of them predicts that the Democrats will have more seats than the Republicans (not counting the tossups). If you average across these different analyses to get a representative prediction, you end up with 216 seats going to the Democrats, 203 going to the Republicans, and 16 being too close to call. If you assign 8 of those tossup seats to the Republicans and 8 to the Democrats, you end up with a final projected total of 224 Democratic seats and 211 Republican seats. Of course, we don't know if the tossups will distribute themselves evenly across the two parties, but that's the best guess as of now. So, at this point, it looks like the Democrats are going to take the House. These analyses fit with the betting at Tradesports.com, which currently gives the Republicans only a 36% chance of retaining control of the House.

But it doesn't fit with what Karl Rove is saying, and I find that to be extremely interesting. It is not surprising to me that President Bush would predict that the Republicans will hold on to both houses of congress. He is, after all, the cheerleader-in-chief, and you'd expect him to say that no matter what. Bush is not claiming to have access to special polling information when he makes that prediction. He's just expressing confidence and exuding optimism to keep the base as fired up as possible.

But it's not that way with Karl Rove. He is claiming to have access to special information that supports his predictions. Here's what he said during a recent NPR interview:

Rove Sees No GOP Fall in the 2006 Election

"I see several things," Rove says. "I'm allowed to see the polls on the individual races. And after all, this does come down to individual contests between individual candidates."

In addition to polls not generally available to the public, Rove says the Republicans have a huge financial advantage in the home stretch.

This seems like more than just cheerleading to me. He is saying that the polling data that we mere mortals do not have access to tell a story that differs from the conventional wisdom. I hope he's right, but I don't believe it! What makes this so interesting for me is that we are going to find out one way or another in less than two weeks. I can hardly wait because I really want to know how predictable elections actually are.

Related to this were these comments that I found at the NRO Corner summarizing a discussion on Fox News:

Fox News All Stars [NRO Staff]
A roundup of last night's panel on Special Report with Brit Hume:

BARNES: Polls don't tell you much about turnout. They really don't and so if you have a great turnout operation, you can thwart the polls.

MARA LIASSON: We know Democrats are more enthusiastic, we know Republicans can get out the vote and the Washington Post poll of Independents show that twice as many Independents are being contacted by the Republicans than the Democrats, that tells you something.

Here again, an informed individual who does not seem to be in cheerleading mode (Fred Barnes) says that turnout operations can thwart polls. Is that really true? I would have thought that both sides would manage to turn out the vote (i.e., it's a wash), which is why averaged polls have been so predictive in the past. But the comments by Mara Liasson point to some objective information suggesting that the Republican get-out-the-vote effort is going rather well, and maybe it will prove the polls wrong. I have trouble believing it, but I'll be tickled pink if it happens.

Maliki's Impossible Choice

I've been arguing for some time now that Prime Minister Maliki won't disarm Muqtada al Sadr's Mahdi Army because the Sunni insurgency wants him to do just that. Maliki will not give the Sunnis what they want until they give him what he wants, which is an end to the insurgency.

While the Sunnis flail away with car bombs in Baghdad (killing some innocent Shiite women and children at a market), the Mahdi Army ruthlessly tortures and then murders dozens of young Sunni men at a time. It appears to me to be an extremely one-sided sectarian conflict, with potential members of the Sunni resistance -- namely, young men -- being systematically exterminated.

For the first time, I encountered someone who seems to share my theory. Fredrick Turner, in an article called Baghdad Vigilantes and the Dark Side of Civil Society has this interesting analysis:

The enemies of the legitimate Iraqi government are almost exclusively Sunnis (though so too are many of its supporters). By logical necessity the exterminators would have to be mostly Shiites and Kurds. The government simply cannot afford to go after its enemies in the systematic way required, for that would be to destroy its claim to represent all minorities. There are, from the point of view of Iraq's nascent civil society, some thousands of people who, in the Texas phrase, need killing. Who is going to do it?

In the absence of government intervention, the answer is: ordinary people. Basically the killers are posses of self-organized vigilantes, who know their local area, who know who the bombers are, and who the bombers' relatives are. The posses are expert in distinguishing those people who might be fair political enemies from those who will go on striking, like a snake, even when cut in two.
...
But death squads are rational, in their own horrible way. They may prove, as they did in Latin America, to be a pretty effective method of wiping out implacable enemies of social order and preparing the way for democratic and law-abiding government. In living memory almost every decent and legal regime in Latin America was preceded by a chaotic period in which ordinary men armed themselves with guns, said goodnight to their families, and went out in groups to kill some local dissident. That period was a bit further back in the past for the French, the English, and the Americans. But no nation can be shown to have reached the rule of democratic law without it. The work of the vigilantes is the hideous and dark crime that Socrates and the Greek tragic dramatists hinted must underlie all civilization. That crime is indeed a crime, and its perpetrators must stand trial for it, whether before God or some human tribunal. But it is possible that true civil self-government can only be established with its aid.
...
But, as Socrates knew, this dark archetypal crime must be hidden. The American authorities in Baghdad are not saying much about it because the vigilantes are doing their work for them, with infinitely greater precision and expertise. The Iraqi government is not doing or saying much about it either, because it would lose legitimacy if it cooperated with the death squads, and sabotage its own interests if it tried (probably unsuccessfully, anyway) to stop them; but it obviously cannot admit that this inaction is its policy. The U.S. Republican Press cannot say anything about it because it would imply in an election year that it approved of the death squads. The Democratic Press cannot give the vital information—that most of the victims probably deserve their fate—because that would imply that the Iraqis have finally started to do what they were expected to do all along, that is, clean up their own house.

It looks that way to me, too. It's unfortunate that it had to come to this, but it is possible that there was simply no way to avoid it. This way of thinking contrasts with what is probably a more common way of thinking, which is captured by a paragraph from this article:

One of the main challenges now is to reassure the Sunnis that it is safe to compromise with Shiite and Kurdish leaders on issues such as the distribution of oil revenue and the shape of Iraqi federalism. Mr. Maliki must also demobilize--or at least neutralize--the militias that grew in his own Shiite community in response to Sunni violence.

That would be the way to go if the Sunnis were capable of being reassured enough to stop the insurgency. Maliki may now believe that no amount negotiating will change Sunni minds unless he keeps his strongest bargaining chip in play. And that, unfortunately, may be the Mahdi Army.

If you could, would you disarm the Mahdi Army if doing so would strengthen the Sunni insurgency and ensure that it continued for years to come? Or would you let that army wear down the Sunni resistance through its mass murder campaign? That may the impossible choice that confronts Prime Minister Maliki. I'm glad I'm not in his shoes.

UPDATE:

Ralph Peters has a new column entitled Kill Muqtada Now. Here's the most relevant excerpt:

I lost faith in our engagement in Iraq last week. I can pinpoint the moment. It came when I heard that Maliki had demanded - successfully - that our military release a just-captured deputy of Muqtada al-Sadr who was running death squads.

As a former intelligence officer, that told me two things: First, Iraq's prime minister is betting on Muqtada to prevail, not us. Second, Muqtada, not the Grand Ayatollah Ali al-Sistani, is now the most powerful man in Iraq.

It's odd to me that Peters does not even consider the relationship between Muqtada al Sadr's Mahdi Army and the Sunni insurgency. They are not completely unrelated issues, and when you consider them simultaneously, what's happening in Iraq can be understood in a different light. It seems far more likely to me that Maliki is using al Sadr's murderous brutality to help fight the insurgency than he is betting that al Sadr will eventually prevail as the new leader of Iraq.

It's the 6th-Year Itch

Not many people are talking about it right now, but the upcoming election looks a lot more like the 6th-year itch than it does a referendum on the Republican vs. Democratic views on Iraq.

First, everyone agrees that the sectarian violence that has generated such pessimism over the future of Iraq jumped to new levels after the bombing of a holy mosque in Samarra back in February of 2006. It's gotten even worse over the past month. Despite that, the proportion of people saying that they will vote for a Democrat in their congressional district has remained pretty much unchanged since way back in August of 2005 (well before the increase in violence). The spike in violence beginning in March of 2006 had a transient effect in favor of the Democrats (as did the Mark Foley micro-scandal in early October), but it is mostly a stable picture:


Second, one Senate race is very clearly a referendum on the Iraq war: Joe Lieberman vs. Ned Lamont. Ned Lamont beat Lieberman in the Democratic primary by running on this one issue, and polls initially showed them to be running fairly even in the general election. Understandably, Lamont is still running on the Iraq issue as things have taken a turn for the worse in Baghdad. Even so, support for the pro-war Lieberman has steadily increased to the point where he now has a commanding lead. That's an odd trend in an election that is primarily a referendum on Iraq.

Third, the outcome of the coming election, if polls are predictive 2 weeks out, will be utterly typical for a 6th-year president. It's called the 6th-year itch (a phrase people generally attribute to Kevin Phillips). The average loss in the Senate for a president's party in the 6th-year of his term is 6, and no modern president has gained a Senate seat in his 6th-year. As such, a loss of 6 seats is the baseline expectation. The Republicans seem on track to lose somewhere between 4 and 6 Senate seats. If this election were a major referendum on Democratic vs. Republican views on Iraq, it would be natural to expect that the President's party would be on track to lose more than the typical number of Senate seats.

Finally, further evidence of a 6th-year itch comes from polls that suggest -- bizarrely -- that the public trusts the Democrats to do a better job with the economy than the Republicans have. Our economy is simply as good as it gets in virtually every measurable respect, and the Democrats have no useful economic ideas other than to repeal the Bush tax cuts that brought the economy back from the brink of disaster after the dot-com crash and the attacks of 9/11. Polls also show that the public trusts the Democrats to do a better job of handling Iraq despite the fact that if you walked into a Democratic caucus on that subject you'd be entering an idea-free zone. What the Democrats have to offer on Iraq is nothing more creative than an involuntary withdrawal reflex. All of this is what you'd expect if this were the typical 6th-year itch (i.e., you'd expect blind trust in the party not in power).

Here is one discussion about this phenomenon from a while back:

This being the "sixth-year itch" election, Democrats could win hands down. For nearly one hundred twenty five years, in the sixth-year of an eight-year presidential term, the party in the White House usually has lost elections. Even Democratic President Franklin Delano Roosevelt (FDR) had election losses. Congress went from 333 Democrats, 89 Republicans and 13 Independents to 262 Democrats, 169 Republicans and 4 Independents in that sixth-year itch election.
...
The sixth-year itch occurred when Reagan lost control of the Senate in 1996. Bill Clinton defied the sixth-year itch by gaining five House seats. That happened because Clinton had taken a very unusual beating two years into his Presidency in 1994 and Republicans maintained control of the Congress in 1996. Even with the sixth-year itch Republicans held on to Congress in 1998.

And here is a table showing how president's have fared in the 6th-year of their terms (numbers found here and here):


The average loss is 36 in the House and 6 in the Senate. Right now, odds are that Bush will do about the same or a little better than that. All in all, it looks like a standard 6th-year itch. Things are rough in Iraq, and that does have people down in the dumps, but, as I see it, the election is not an endorsement of what the Democrats have to offer. It's more like blind faith in the only real alternative we have. And that's what I mean when I say that the upcoming election -- if it turns out as predicted -- will best be viewed a classic example of the 6th-year itch.

Two Weeks To Go

There is a lot of talk in the blogosphere about the House and Senate races improving for the Republicans. For example, John Podhoretz says:

By last Thursday, however, those of you on Blog Time began to discern a change: Suddenly, things weren't quite so bad for Republicans or quite so great for Democrats.

Blog Timers adduced this not from major evidence, like big polls, but from small data points - trees rather than forests.

Blog Timers noted a Maryland poll indicating a tie in the Senate race there, a seat Democrats are counting on. In Tennessee, they saw Democratic Senate candidate Harold Ford deciding to confront his GOP rival, Bob Corker, at a Corker event - and for the first time in a long and confident campaign, looking desperate and worried.

And here's Dick Morris and Eileen McGann saying pretty much the same thing:

October 24, 2006 -- THE latest polls show something very strange and quite encouraging is happening: The Republican base seems to be coming back home. This trend, only vaguely and dimly emerging from a variety of polls, suggests that a trend may be afoot that would deny the Democrats control of the House and the Senate.

Well, I wish I could see what everyone else does, but I just don't (at least not with regard to the critical Senate races). What jumps out at me is the absolute stability of the numbers. Here is how things stand with 2 weeks to go:


Again, the poll results are not from any one poll (because, contrary to what almost everyone seems to believe, individual polls are barely informative) but they represent the average of multiple polls (because averaged polls are much more informative). The first 7 states shown are the ones where a party switch might happen. The last 2 states -- Connecticut and Virginia -- are also shown because there as been a lot of interest in those races, though a switch from one party to the other seems unlikely in those cases.

The states shown in red are currently held by Republicans and those shown in blue are currently held by Democrats. The story told by the averaged polls is the same as it has been for several weeks now: the Democrats are on track to hold on to New Jersey and to pick up 4 more seats (Montana, Ohio, Pennsylvania, and Rhode Island). Another two races are tossups (Missouri and Tennessee). The betting on the individual Senate races at Tradesports.com (also shown in the table above) tells much the same story, though things appear to have tilted slightly in favor of the Tennessee Republican (Bob Corker) lately.

The overall betting at Tradesports.com has remained mostly unchanged as well: there is a 67% chance that the Republicans will hang on to the Senate and a 35% chance that they will hang onto the House. That's pretty close to how I see it as well (though I put the chances of retaining the Senate at 75%).

All of this also fits with Gallup's generic poll question that asks people which party's congressional candidate they will vote for on Nov. 7. The same story has emerged from that question for a long time (favoring Democrats):


You can see what looks like an abberrant tie in mid-September (47% vs. 45%) and then an aberrently large tilt in favor of the Democrats shortly thereafter (probably due to the Mark Foley micro-scandal), but it's basically a stable story: people generally say they'll vote for Democrats (about 53% on average) over Republicans (around 40%). Note that these numbers have not really changed much with the worsening of the situation in Iraq, which is one of many reasons why I do not see the upcoming election as a referendum on Iraq. Instead, it would appear to be the standard 6th-year itch that has plaugued every president in the 6th-year of his term (as I discussed here).

Don't get me wrong. I feel that the Republicans will hold on to both houses because I simply cannot believe that Americans want what the Democrats have to offer. I reviewed what the Democrats are likely to do if they win here, and it isn't a pretty picture even when told by reporters who are sympathetic to the left. Pete DuPont has a new article on this topic as well in which he asks Republicans deserve to lose, but what happens if Democrats win? Here are some answers:

First, Minority Leader Nancy Pelosi has promised that election of a Democratic House would insure "a rollback of the [Bush] tax cuts." Rep. Charles Rangel of New York, who would be chairman of the Ways and Means Committee, would make sure no tax cut extension bill would ever get to the floor. He voted against the 2001 and 2003 Bush tax cuts and the bill that later extended the tax cuts until 2010 (as did all but seven of the 205 Democratic House members). In September Mr. Rangel said that he "cannot think of one" Bush tax cut he would agree to renew.
...
As for the war in Iraq, Mr. Rangel observed that "You've got to be able to pay for the war, don't you?" In other words, end it by simply defunding it. Rep. John Murtha of Pennsylvania calls for "immediate redeployment of U.S. troops" and intends to run for majority leader if the Democrats take control of the House. Ninety percent of House Democrats opposed the terrorist surveillance program, and 80% voted against the recent terrorist interrogation legislation.

Finally, when we see what the new leaders of a Democratic House are likely to do, their views are--well--very different from most Americans. Rep Henry Waxman of California would become the Government Reform Committee chairman, and believes domestic terrorist surveillance is "illegal." He would use his subpoena power to launch investigations to try and limit the president's anti-terrorism powers.

Rep. John Conyers of Michigan, who would become chairman of the Judiciary Committee, has talked about subpoenaing "Bush administration officials to answer questions and face the consequences for their abuses of power." In other words, impeachment.

Speaker-to-be Nancy Pelosi has indicated she would like to put Rep. Alcee Hastings of Florida in charge of the House Intelligence Committee. As a federal judge, he was impeached in the House by a 413-3 vote, and removed from the bench by the Senate for bribery, corruption, and perjury. Rep. Hastings would lead the oversight of America's antiterrorism policies.

All of this and no clue about what to do in Iraq beyond a primal instinct to withdraw (despite the horrendous consequences if we do). It's hard to imagine America putting these keystone cops in charge, but, right now, the evidence suggests that we will.

The Republicans have more money going into the home stretch, and one can only hope that commercials showing Pelosi, Rangel, Conyers, Murtha and Hastings will shock Americans out of their apparent slumber. In fact Baron's Magazine thinks that the financial advantage enjoyed by the Republicans will, in the end, do the trick:

Our analysis -- based on a race-by-race examination of campaign-finance data -- suggests that the GOP will hang on to both chambers, at least nominally. We expect the Republican majority in the House to fall by eight seats, to 224 of the chamber's 435. At the very worst, our analysis suggests, the party's loss could be as large as 14 seats, leaving a one-seat majority. But that is still a far cry from the 20-seat loss some are predicting. In the Senate, with 100 seats, we see the GOP winding up with 52, down three.

We can hope, and we'll know in two weeks time. I can hardly wait to find out what the best predictors are this time around. For now, though, I see the Democrats taking the House and coming close to taking the Senate.

"Stay the Course"

I am amazed by all the fuss over the phrase "Stay the course" in Iraq. Last night on Air America, Randi Rhodes was gleefully pouncing on it, as was Al Franken this morning. Bush has decided not to use the phrase anymore, and others have said that the White House never had a "stay the course" policy. This is what the left is positively gleeful about because they have Bush on tape using that exact phrase quite a few times (and they play that tape over and over again). Al Franken and his co-host were laughing hysterically at the apparent contradiction as the tapes were played.

I know that Al Franken is an intelligent individual, but he was not acting like one this morning. Bush explained almost a year ago what he means by that phrase and how others have a habit of misinterpreting it:

"If by stay the course, they mean we will not allow the terrorists to break our will, they are right. If by stay the course, they mean we will not permit al Qaeda to turn Iraq into what Afghanistan was under the Taliban -- a safe haven for terrorism and a launching pad for attacks on America -- they are right, as well," the president said. "If by stay the course, they mean that we're not learning from our experiences, or adjusting our tactics to meet the challenges on the ground, then they're flat wrong."

Obviously, the policy never was to "stay the course" in the sense that liberals have chosen to misleadingly interpret that phrase. And that's what Bush and others mean when they say that they never had a "stay the course" policy (i.e., they never had a policy to do what liberals assume when they hear that phrase). No one is claiming that Bush never uttered those words, so playing a tape of him doing so is to badly miss the point. Bush obviously uttered those words, but -- just as obviously -- he meant something different from what the left imagines.

Even though Bush won't use the phrase anymore, it's worth reminding people what it really means and what the choices are:

Stay the course = we will not allow the terrorists to break our will

Change the course = we will allow the terrorists to break our will

Take your pick. While others fuss over the phrase itself, a more substantive debate could be had by focusing on its actual meaning as intended by the president. I hope that Air America (and the lefty blogosphere) will step up to the plate once they have come to appreciate their mistake. Or might they be making that mistake on purpose?

Dear CNN: Here is How to Handle Propaganda Films Sent to You by the Enemy

The Washington Post just published an article on the CNN sniper video controversy. They quote Rep. Duncan Hunter, chairman of the House Armed Services Committee, as saying that CNN's airing of the video was irresponsible because it could encourage more attacks on U.S. troops. And he added:

"Does CNN want America to win this thing?" Hunter asked yesterday on the network. In past wars, he said, the media was more pro-American.

"You can't be on both sides of the war," Hunter said.

The article then presents CNN's side of the story:

CNN issued a statement saying that the decision was "a difficult one, but for a news organization, the right one. Our responsibility is to report the news." The video was delivered to the network through a contact with an insurgent leader.

No, airing the tape was the wrong decision. Could it be that CNN simply doesn't know what the right decision would look like? Allow me:

CNN has come into possession of a chilling video sent to us by an insurgent leader in Iraq. The video shows snipers targeting unsuspecting American troops, some of whom are then shot on camera. It seems clear that the tape was sent to us for propaganda purposes, but it also clear that the tape has news value. As such, we have decided to acknowledge the existence of the tape and to describe its contents. However, we have also decided not to show the video -- as the insurgent leaders clearly hoped we would -- because to do so would be to serve the propaganda purposes of the enemy. We hope that readers understand our difficult decision.

We turn now to a more detailed description of the contents of the video. The details may be upsetting to some readers, but we feel that the video contains important information, and we have a responsibility to report the news.

That, CNN, is how you do it. Report the news without also advancing the cause of the enemy. Give it a try the next time the enemy picks you to serve as its American propaganda wing.

What Past Elections Suggest About Our Upcoming Election

I love the fact that we approaching another election because it provides such a great opportunity to evaluate the validity of before-the-fact thinking. As you might already know, I always tell people to ignore individual polls and to focus instead on averaged polls. The reason I do that is because of what I learned the last time we held a national election. For example, using nothing but averaged state polls, RealClearPolitics made this projection on the morning of the last presidential election:

Projected Elector Tally: Bush 296 - Kerry 242

And here is how the election actually turned out:


The projection missed only in Wisconsin, so their 10 electoral votes went to Kerry instead of Bush. But that was good enough for me. In the end, the outcome of that seemingly very uncertain race was, in fact, quite predictable. I remember hearing about the exit polls on election day and thinking, somewhat to my surprise, that the pre-election polls were all wrong because Kerry had just been elected. As it turned out, the pre-election polls were right on the money (once they were averaged) and the exit polls were pretty much useless.

The average of the national polls predicting the percentage of the popular vote turned out to be pretty accurate as well:


The final projected tally (Bush = 48.9%, Kerry = 47.4%) gave you a pretty good idea of what the final voting result would be (Bush = 50.7%, Kerry = 48.3%). Because many polls have been averaged together to create this chart, you can see meaningful trends (i.e., the trends represent more than just statistical noise). The 2004 Democratic National Convention was held at the end of July, and you can see the effect it had on the race, with Kerry jumping into a clear lead at the beginning of August. The 2004 Republican National Convention was held August 30 - September 2. You can see that Kerry's lead was already evaporating by that time (the Switft Boat ads were having an effect, I believe), and then the Republican convention had a large effect in favor of Bush that lasted until the election was held.

Today, the average polls suggest that the Democrats will pick up several seats in the Senate, and it is hard to imagine that it won't happen. After all, it fits with what one would expect even if we had nothing to rely on but historical trends, as Larry Sabato pointed out a while ago:

Let's look at the last half-century of midterm elections for the U.S. Senate from 1950 to 2002...On average, the president's party has lost three Senate seats in each of those 14 elections.

So that's the baseline expectation. In fact, it's even worse than that because of what might be called the 6th-year itch (i.e., anti-incumbent results that occur in the 6th-year of a president's term):

There have been six of these elections in the post-World War II era (1950, 1958, 1966, 1974, 1986, and 1998). The average loss for the White House in these sixth year elections has been six Senate seats--double the overall midterm average loss of three seats.
...
Never in modern times has a president been able to add Senate seats in the dreaded sixth-year election.

In other words, at best, it is tough going for a president in his 6th-year. According to my last look at the averaged polls, the Republicans will lose 5 plus-or-minus 1 Senate seats on November 7, in which case this would, at worst, turn out to be an absolutely typical 6th-year election.

It is interesting to think of the election in that light. That is, it may not really be such a big referendum on Bush's policy with regard to Iraq. Instead, it's just the typical 6th-year itch. This doesn't mean that people are happy with the way Bush is handling Iraq, but it does mean that the issue is not driving people in droves to vote for Democrats (at least not more so than would be typical in a 6th-year election). This interpretation fits with the fact that the one Senate race that is very definitely a referendum on the Iraq war -- Joe Lieberman vs. Ned Lamont -- is turning into a rout in favor of the hawk (not the dove). Here are the averaged poll results for that race:


As you can see, Lieberman is ahead by a whopping 12.5% right now, which means that he is very likely to be reelected. It seems unlikely to me that it would be that way if there were a big anti-Iraq political sentiment sweeping the nation. What we are going to see on November 7 is the ordinary 6-year itch, not an extraordinary referendum on Bush's policy in Iraq.

Insights Into What Victorious Democrats Might Do

I just read two articles about what the Democrats are likely to do if they regain power in the upcoming election. The first article is this one from Newsweek magazine:

What the Dems Would Do
They've waited 12 long years to reclaim the steering wheel. How the party out of power would rule if they retake the House

by Michael Isikoff and Holly Bailey (with Mark Hosenball and Eleanor Clift)

These are left-leaning reporters (I believe), so it is presumably the positive spin on what would happen if the Democrats win the House and Senate. The second article (via RealClearPolitics) is:

If the Democrats take Congress, here's what they'd do

by David S. Broder

He, too, is a left-leaning columnist, so, again, it's the rosy view.

As you might expect, neither article tries to claim that the Democrats will straighten out the mess the Republicans have made of the American economy. Our economy is so undeniably fabulous that any suggestion that the Democrats might do a better job in that regard would be impossible to take seriously. Both articles instead focus on a traditional Democratic weakness: national security. Supposedly, the Democrats will straighten out everything the Republicans have done wrong in that domain.

And how will they do it? In their article, Isikoff and Bailey note that the Democrats will focus on domestic wiretapping, port security, and the like, which is fine, but one wonders what they will do about the most important national security issue of our time, namely, Iraq. On this point, Isikoff and Bailey do not mince words:

Democratic leaders are largely sticking with domestic issues in part because they have yet to come up with a coherent plan for the biggest problem of all: articulating a clear way out of Iraq. On the campaign trail, Democrats have been content to bash Republican failures and say they'd do better. The Democrats' official line is to promise a "new direction," and to urge, vaguely, "redeployment." If they win, they'll be forced to say what, if anything, that means.

Don't you think they should be forced to say what, if anything, that means before the election? It seems clear that the Democrats have no plan at all beyond a basic instinct to withdraw. Perhaps David Broder's article can help to clarify the Democratic plan for Iraq:

WASHINGTON — No one speaks more authoritatively for the Democrats on defense and national-security issues than Sens. Carl Levin of Michigan and Jack Reed of Rhode Island, both longtime members of the Armed Services Committee. If you want to know what Democratic gains in this midterm election would mean for national security policy, Levin and Reed can provide the answers.

Excellent. From these two big wheels we can finally get the authoritative plan of action that the Democrats have devised for Iraq. And here it is:

On Iraq, the two Democrats harked back to the amendment that 39 senators supported during a debate earlier this year — an amendment that called for a start on U.S. troop withdrawals within six months, but set no numbers and specified no target date for ending the U.S. military presence.

Reed, who has made many trips to Iraq and returned just weeks ago from his most recent visit, described the "very, very difficult situation" he found there. "We have to begin to work toward redeployment without setting a timetable," he said. "We have to start laying out some red lines for the Iraqis ... give them some clear goals we want them to achieve."

Did he really say "We have to begin to work toward redeployment without setting a timetable"?

Now you see what Isakoff and Bailey were talking about when they said: The Democrats' official line is to promise a "new direction," and to urge, vaguely, "redeployment." If they win, they'll be forced to say what, if anything, that means.

As you can plainly see, it doesn't mean anything. It's just a trick to get people to vote for Democrats. You don't need to be a rocket scientist to know that the competing plans are to (a) redeploy when the Iraqis can manage their own security (the Republican plan) or (b) redeploy according to a timetable even if the Iraqis are not able to manage their own security (the mostly unspoken Democratic plan). Reed preposterously tries to straddle these alternatives, presumably to disguise the fact that he favors the second option. But no matter what the Democrats say, these are the only two options. Take your pick, and then vote for a Republican or a Democrat (depending on which plan makes more sense to you).

Right now, most people seem inclined to vote for a Democrat. Well, from the Newsweek article, at least we'll have these changes to look forward to:

If Democrats take it back next month, the party will once again be in charge of all the committees. Dingell—now 80 years old and more ornery than ever—is all but certain to return to his old job [chairman of the House Energy and Commerce Committee]. After 12 long, frustrating years as the panel's ranking minority member, a title that left him little more than the power to complain, nothing animates Dingell more than the thought of making up for lost time.
...
Henry Waxman of California is another Democratic old-timer whose ire never dims. A tireless investigator, he's in line to head the Government Reform Committee, and plans to take aim at Halliburton and alleged rip-offs and contract abuse in Iraq. Then there's Charles Rangel, the New York congressman who's never met a cable show he didn't like. He is set to take over the Ways and Means Committee, and wants to take a hard look at the Bush tax cuts. John Conyers of Michigan has waited for years to head the Judiciary Committee. He's likely to convene hearings on the Patriot Act and domestic wiretapping. In the past, he has suggested the possibility of impeachment hearings for President Bush.

As I said, this is optimistic spin provided by reporters who sympathize with the liberal agenda. As such, this is the rosy view.

Good heavens, America. What are you thinking?

The Options in Iraq

A few days ago, I stumbled across an interesting and illustrative exchange that seems to perfectly capture the differing views on Iraq:

Foreign Minister Alexander Downer, in parliament yesterday, clashes with the Labor Opposition over Iraq

Opposition defence spokesman Robert McClelland: Does the Foreign Minister recall saying on July 5, 2004, that the Iraqis were making "sound progress"? Does he recall saying on September 19, 2005, that the situation in Iraq was "headed in the right direction"? Does the Foreign Minister acknowledge that, with 100 civilians dying every day in Iraq, Australia is now effectively bogged down in a full-blown civil war? As the minister and the Prime Minister got Australia into this hole, when will they stop digging?

Downer: The answer is a series of yeses and noes, but I make this point in answer to the last part of the question: we will, along with our allies, pass the responsibility for security in Iraq over to the Iraqi security forces when they have the capacity to maintain a reasonable degree of security in that country. The Labor Party strategy is for all of the coalition forces to withdraw from Iraq and to create not only a human rights catastrophe but also an enormous boost to the extremists and insurgents.

Doesn't that discussion nicely encapsulate the liberal and conservative positions on Iraq? If you are a liberal opponent of the war in Iraq, you probably love to complain about the war in just the same way that Robert McClelland does. Complaining about what Bush has done (or is doing) in Iraq is the essence of the liberal point of view on that subject. Although you might resonate to that, I resonate to Downer's response instead, which focuses on what to do now and on the consequences of doing what his liberal opponent apparently wants to do (but will not actually say). Complaining about what we are doing in Iraq is both easy and useless, but, for the most part, it's what liberals do.

On a related note, a recent story in the Washington Post says that the Bush administration will soon be forced to consider new options in Iraq. I applaud the reporters (Michael Abramowitz and Thomas E. Ricks) for doing more than complaining and for actually discussing what should be done. We need more of that from the angry left. Here are the options that they believe will soon be recommended:

One would be some kind of effort to divide the country along regional lines. Another, favored by many Democrats, is a gradual withdrawal of troops over a set period of time. A third would be a dramatic scaling-back of U.S. ambitions in Iraq, giving up on democracy and focusing only on stability.

Unfortunately, none of these sound like reasonable ideas. Option #1 (segregating Iraq) has this little problem:

New Poll Finds Iraqis Favor Unity
Overwhelming Majority Oppose the Idea of Segregation
FOR IMMEDIATE RELEASE
July 19, 2006

Washington, D.C. --A new poll released by the International Republican Institute (IRI) found Iraqis strongly oppose the idea of segregating the country. Seventy-eight percent strongly disagree or disagree with the idea of segregating Iraqis according to religious or ethnic sects. An overwhelming majority, 89 percent, believe that a unity government is extremely important to Iraq's future.

Here are the results of that poll broken down according to religious and ethnic group (red and orange indicate opposition):


As you can see, only the Kurds favor the plan, and even they are not wildly enthusiastic about it. I feel pretty sure that Turkey would be adamantly opposed as well. And what would you do with the 2 million Sunnis living Baghdad? Forcibly evict them from their homes in the same way that Muqtada al Sadr's militia is today? I don't see how this plan can work, but I was nevertheless glad to see it adopted by Air America the other night. I could hardly believe that Randi Rhodes stopped her Bush bashing long enough to endorse a specific plan (misguided though it may be). That's not something that happens every day.

Option #2 (the Democratic plan to set a timetable for withdrawal) has the problem that it will, according to the recent NIE report (and according to common sense) be perceived as a grand victory by al Qaeda, which will greatly strengthen the jihadist cause around the world. It will also ensure that we leave Iraq a failed state, which will further ensure that it will evolve into the perfect breeding ground for international terrorism (think Afghanistan under the Taliban). I know that critics believe that we are creating more terrorists now, and we surely are (4000 of them are dead now, in fact, having been predictably drawn to Iraq like flies), but the ultimate effect on the global jihadist movement hinges on what we do now. If we pull out and leave Iraq a failed state (i.e., if we implement option #2), we will not only dramatically accelerate the creation of new jihadists emboldened by their glorious victory over America (according to the NIE report), we will also give them the perfect place to conduct their business. And since we won't be confronting them there anymore, they will be freed up to attack American interests around the world. As a purely emotional reaction, a preference for Option #2 is understandable. As an actual course of action, it is self-defeating.

And I just do not know what to make of the 3rd option. Give up on Democracy? What does that mean? Announce that we are canceling further elections and installing Maliki or someone else as the new dictator of Iraq? Should we support the use of genocide to get matters back under control as well?

These ideas don't seem sound to me, but it's nice to see a liberal newspaper discussing ideas instead of standing on the sidelines shouting "you're doing it wrong!" That's pretty much all that defense spokesman Robert McClelland did in the debate linked to above, and it's what most Democrats running for office are doing as well. I hope that by election day, House and Senate candidates will take a stand on what they think we should do about Iraq. But because that would almost ensure a Republican victory at the polls, I sincerely doubt it will happen. Americans seem poised to give some power to the Democrats even though their position on Iraq amounts to little more than shouting "you're doing it wrong!" That being the case, expect the Democrats to shout louder instead of clarifying their position on the most important matter of our time.

"The right picture can win or lose a war"

In Clint Eastwood's new movie Flags of Our Fathers, there is an observation that caught my attention in light of CNN's decision to air a sniper attack video sent to them (to them in particular) by our enemy in Iraq. A review of the movie says:

We also hear photographer Rosenthal as he attempts to explain why his picture touched a national nerve. "What we do in war, the cruelty is almost incomprehensible," he says. "But somehow we need to make sense of it. The right picture can win or lose a war. I took a lot of other pictures that day, but none of them made a difference. Looking it at, you could believe the sacrifice was not a waste."

The line that caught my attention was this:

"The right picture can win or lose a war"

Here is the photo that once helped us win a war:


The photograph was used as propaganda, to be sure, but those men were heroes, and what they did was inspiring. That picture helped to inspire an entire nation -- our nation.

But that was then; this is now:



These are screenshots from the CNN sniper video. The rest of the video, showing American soldiers being shot by Sunni insurgents, is every bit as demoralizing as it was intended to be.

CNN's actions have elicited a bit of a backlash:

SAN DIEGO The chair of the House Armed Services Committee asked the Pentagon today to remove C-N-N reporters embedded with U-S combat units.
...
Representative Duncan Hunter wrote in his letter that, quote, "C-N-N has now served as the publicist for an enemy propaganda film featuring the killing of an American soldier."
...
C-N-N executives defended their decision to air the footage, saying its news value outweighed other concerns.

The news value of the video outweighed other concerns? To them, I'm sure it did. One wonders if, in the minds of CNN executives, the famous Rosenthal photograph depicting the inspiring bravery of the American soldier would have had news value that outweighed other concerns as well. I somehow doubt it. After all, it wasn't that long ago that this happened:

CNN executive resigns after controversial remarks
Friday, February 11, 2005 Posted: 10:14 PM EST (0314 GMT)

ATLANTA, Georgia (CNN) -- CNN chief news executive Eason Jordan resigned Friday, saying the controversy over his remarks about the deaths of journalists in Iraq threatened to tarnish the network he helped build.

Jordan conceded that his remarks at the January 27 World Economic Forum were "not as clear as they should have been." Several participants at the event said Jordan told the audience U.S. forces had deliberately targeted journalists -- a charge he denied.

CNN has a problem, and it's clearly a bad problem. They are so infuriated with George Bush and so dedicated to tearing him down that they can no longer tell when their actions do more to advance the cause of the enemy than they do to inform their audience. The right picture can win or lose a war, and I wish CNN would decide which way they'd rather see this war go. Perhaps they already have.

The Mahdi Army is Bad for the Sunni Insurgency

Bill Roggio (via Instapundit) says that we need to crack down on Muqtada al Sadr's militia, and he says:

Sadr apologists have repeatedly claimed that elements of his Mahdi Army have begun to operate outside of his control, however Sadr has done nothing to rein in these "rogue elements," except making vague statements about halting the sectarian violence. He has not identified these out of control militias, closed offices or denounced them specifically. Sadr is playing a double game of maintaining his militia against the law while pretending to be a responsible member of government.

I mostly agree, except for the "pretending" part. My theory is that his militia is playing a strategic role in the fight against the Sunni Insurgency, which is the main reason why Maliki is reluctant to finally confront him. That is, it's not just that Maliki is politically dependent on the support of al Sadr. Instead (I theorize), Maliki knows that the Mahdi Army is, by far, the most serious threat to the intransigent Sunni population in Iraq and, indirectly, to the Sunni insurgency. The Sunni population has been uniformly hostile to the changes that have been forced upon them since we toppled Saddam Hussein, and they have been especially hostile to the presence of American troops in their country (hence, their support for the Sunni insurgency). According to polls from Iraq, their attitudes have remained utterly hostile since day one.

Until recently, that is. As I noted in a prior post, recent polls show that:

...the Sunnis’ earlier overwhelming eagerness for withdrawal [of US troops] has moderated, with the percentage calling for withdrawal within six months dropping from 83 percent to 57 percent. Among those living in Baghdad support is even lower at 24 percent.

The Sunni insurgents, like the larger Sunni population, cannot be cheered by the fact that the Mahdi Army systematically executes relatively young Sunni males by the dozens every few days. As such, unilaterally dispensing with the Mahdi army would greatly strengthen the hand of the Sunni insurgency, and that, I believe, is a factor is this apparent waiting game.

Consider this recent interview with Prime Minister Maliki. He was asked "When will the militas be disbanded?" His response:

Regarding setting a time, I don't think we could determine it specifically. The problem of militias, in countries throughout the world, requires time. The most important thing is that we have started and started strong. We have given a clear message: Militias should reconsider their existence. ... The more success we have on the political side will help us deal with this issue. The initial date we've set for disbanding the militias is the end of this year or the beginning of next year.

What is he waiting for? The Mahdi Army poses many problems for the Iraqi government (e.g., they infiltrate the police, they impose Islamic law where they are in control, etc.). But al Sadr's army is also bad for the Sunni insurgency. That's why the Mahdi Army still roams relatively freely (according to my theory, anyway). Maliki may be saying to the Sunnis (in effect) "we will not do you the great favor of doing away with the Mahdi Army until you do me the great favor of giving up the insugency." That may be the high-stakes scenario that we are watching unfold in Iraq.

The interviewer at USA today seems to share my theory (note Maliki's vague response):

Q: U.S. military leaders have said it's the militias, specifically the Mahdi Army, doing most of the killing. Why not confront them now? Is part of the strategy to let the militias operate as long as there's an insurgency?

A: The military decision is to confront anyone carrying a weapon who is outside the law, whether it's the Mahdi Army, the Omar Army or the Islamic Army in Iraq. All of them will be confronted, whether it's by international forces or Iraqi forces.

We're dealing with a very delicate equation. That's why we are applying military and political pressure at the same time. During the period of time required for this approach, our security forces become stronger and the crimes committed by militias become clearer and clearer to everyone. Then, when we confront the militias, there will be no negative reaction to confronting them, especially from the people.

Previously, when we attacked some of the Sunni death squads and militias, we were being accused of being biased toward Shiites against Sunnis. Or when we confronted the Mahdi Army, we were accused of being biased toward the Sunnis against the Shiites. But now it's becoming clear, after we confronted the Mahdi Army in Diwaniyah, Karbala, Basra and Nasiriyah, no one can say we are biased on this issue.

This is a very important step we have managed to achieve. Now no one can say we are biased when we hit a Sunni militia. So the road is being paved and the forces are being prepared to resolve this issue of militias within the time frame we discussed.

I think a more succinct and much more accurate answer to the interviewer's question would have been "yes."

Finally, we have this from the interview:

Q: Do you feel you're close to a political solution?

A: Yes, definitely.

I suspect that intense negotiations form the backdrop to the bloody chaos we are seeing in Iraq. Maliki thinks that a solution is near, and, if so, my theory is that the brutality of the Mahdi Army is partly responsible for that.

Poll Watching in the Blogosphere

My efforts to convince the commentariat not to overreact to individual poll variability has fallen completely flat. But you, dear reader, will have special knowledge about polls after reading this post.

Most polls of Senate races involve about 650 likely voters. The margin of error in polls of that size is about 4%. Imagine 20 polls showing that Candidate A leads Candidate B by an average of 5%, with some polls suggesting that the lead is as low as 3% and others showing that the lead is as large as 7%.

What would your reaction be? A natural reaction might be:

"Well, the polls are in pretty good agreement, and the range of leads for Candidate A (3% to 7%) is about the same as the margin of error, so I find the results pretty convincing. Candidate A leads by about 5%."

Your reaction instead should be:

"No possible way. These results cannot be true because the amount of variability across polls is way less than the amount of variability that ought to be there."

Now imagine 20 more polls showing that Candidate A leads Candidate B by an average of 5%, with some polls suggesting that Candidate A leads by 1% and others showing that the lead is as large as 9%.

Your reaction to this might be:

"Gee, the polls are flopping around a bit. I can't tell if Candidate A has a large lead or a small lead, but it seems clear that Candidate A is in the lead."

Your reaction instead should be:

"No possible way. These results cannot be true because the amount of variability across polls is way less than the amount of variability that ought to be there."

Finally, imagine 20 more polls showing that Candidate A leads Candidate B by an average of 5%, with some polls suggesting that Candidate B actually leads by 3% and others showing that Candidate A leads by as much as 13%.

Your reaction to this might be:

"This race shows extreme volatility. The numbers are all over the place. I don't know what is going on, and I can't even tell who is in the lead. I'm not sure I even believe polls anymore."

Your reaction instead should be:

"Candidate A is clearly in the lead by about 5%. The variability across polls is right about what you would expect for sample sizes of 625."

Get the picture? Intuition is a poor guide when it comes to interpreting polls. Here is the simple rule (explained in more detail here): take the margin of error and multiply it by 4 (that's right, by 4), then expect the lead to vary in that range. For polls based on 625 people, the margin of error is about 4%. If you multiply 4% by 4 you get 16%. That's the range of leads you should expect (approximately) due to random chance alone.

Let's take a real world example that has a lot of bloggers talking. The latest poll of the Senate in Maryland shows Republican Michael Steele in a sudden tie with Democrat Ben Cardin. Other recent polls have Cardin in the lead by 6, 8, 9 and 15 points. Thus, the lead goes from 0 (a tie) to a 15-point lead for Cardin. The average lead is 7.6%.

Here's my question: do you believe that Cardin and Steele are tied, as the latest poll suggests, or do you believe that Cardin has a 7.6% lead? You should definitely believe the latter because the range of leads (0 to 15 points) is right in line with what you'd expect if the true lead were 7.6% and the sample sizes were in the range of 650.

The mighty Powerline says:

The latest Survey USA poll of the Maryland Senate race shows Michael Steele and Benjamin Cardin tied with 46 percent each. As with the last Survey USA Maryland poll, Republicans may have been oversampled. This time Democrats exceeded Republicans by 18 percentage points (as opposed to 16 points last time). The Democratic lead in voter registration is more like 25 points.

Still, the poll gives Steele reason for hope. And, as I've said before, this race may be particularly volatile and hard reliably to poll because of the racial politics involved.

They are appropriately cautious here, but in response to a later poll showing Cardin in the lead, they add:

These results may not be correct, but almost all of them strike me as plausible. One exception may be Maryland. I think Cardin is ahead, but I question whether he's 8 points ahead.

I don't! The results of multiple polls are right in line with what you'd expect if Cardin has a 7.6% lead. I wish it were otherwise, and perhaps it will be by election day, but right now, the polls are behaving as they should if Cardin has a 7- or 8-point lead.

AJ Strata says:

I predicted last year Republican Michael Steele would win MD’s open Senate seat. The fact is the black community in MD is not going to vote in another lame white guy when they have Steele on the ballot, with the best campaign commercials ever produced. And now we see a poll showing a tie and this race - which means a pick up for Reps (with NJ in most likely)

Not so fast! No one can say for sure, but, for now, the safe bet is that Cardin leads by 7 or 8 points despite the new poll (though I sincerely hope that AJ Strata's prediction turns out to be right). This is why the actual betting at Tradesports.com is 79/28 in favor of Cardin despite the new poll showing them to be tied.

And the always great Jim Geraghty has a post entitled "A review, and the upset special prediction (UPDATE: Steele's tied up with Cardin)." No he isn't! Cardin is up by 7 or 8 points according to the evidence.

If you read my blog, you know the rule, but I shall repeat it anyway: ignore individual polls and let your thinking be guided by averaged polls. Individual polls allow people to make news out of random error, but they don't really tell you anything (well, not much of anything).

Highs and Lows

Worth noting:

Consumer Confidence Reaches New High for 2006 at 120.5

NBC/WSJ poll: Public's opinion of GOP hits record low

OK, it's Iraq. The insurgency there has apparently launched its own version of the Tet Offensive. It worked for the enemy in Vietnam and, if the polls are any indication, it may work for the enemy in Iraq as well. The party in the lead is the party that wants to set a timetable for withdrawal. According to the recent NIE report, a premature withdrawal from Iraq will embolden the terrorist movement around the world. And according to every reasonable analyst out there, leaving Iraq a failed state will virtually ensure that those emboldened terrorists will use that country to advance their jihadist cause.

I have a hard time believing that we will actually turn both houses over to the Democrats in November. Instead, it seems to me that the polls reflect understandable unhappiness with how things are going in Iraq, not an actual preference for the unintentionally pro-terrorist alternative offered by the Democrats. By election day, I expect these matters to come into sharper focus, in which case the situation will be less favorable for Democrats than it is today. But that expectation could be clouded by how I hope things turn out.

Maj. Gen. William Caldwell said:

"The enemy knows killing innocent people and Americans will garner headlines and create a sense of frustration. However, the coalition will not be deterred from establishing an Iraq that can provide for its own security and govern itself," he said.

Oh yes they will be if the Democrats have anything to say about it. And the enemy in Iraq knows it.

Why Does Average Income Go up when Median Income Goes Down?

One view holds that the economy is doing spectacularly well according to any standard measure (GDP growth, unemployment, productivity, federal deficit, stock market, inflation, etc.). The other view holds that while the statistics look good, only the wealthy are enjoying the benefits. Paul Krugman sums up this view perfectly. "In short, it’s a great economy if you’re a high-level corporate executive or someone who owns a lot of stock. For most other Americans, economic growth is a spectator sport." I love that line. Believing it to be true is the hallmark of a liberal mind, and I don't mean that in a bad way. It's good to have people around who are utterly convinced that the little guy is being left behind because such people ensure that we don't overlook any evidence suggesting that it might be true. At some point, though, you'd like the evidence to actually influence how you think about this issue.

The idea that the average fellow is suffering in Bush's economy has reached epidemic proportions. Kevin Drum, for example, says that "...median incomes have dropped even though the economy has been growing" and then adds:

I'll say one thing, though: those boys down in Texas sure did a whole lot better under Clinton than they have under Bush. If they were smart, they would have voted for Gore and kept the Shrub under wraps in Austin, where he couldn't have done so much damage to their economy.

Similarly, the Democratic National Committee has a web page entitled "Bush Economy Fails Working Families." And in an article entitled "The end of the American dream?", the BBC says:

For real household incomes, the median point - the level at which half of households earn more and half less - has actually fallen over the past five years.

That marks a notable contrast with the 1990s, when the economic boom boosted both jobs and incomes.

The main argument in favor of the claim that people in the middle are not benefiting from our fabulous economy comes from income figures supplied by the Census Bureau. Median incomes (i.e., incomes for those in the middle) have, indeed, dropped during the Bush years. What more proof do you need? Here is an income chart that shows this effect (the Bush years are shown in red):


The data were taken from the Census Bureau report, which breaks incomes down according to quintile (quintile 1 = the poorest 20%, quintile 5 = the richest 20%). The graph shows the results from the middle quintile (those in the 40% to 60% range), which is just like median income. Sure enough, incomes have not done very well since Bush took office according to the Census Bureau measures, and some believe that this explains why people are down on the economy in spite of the fact that all of the other statistics suggest that things are going very well. If you are not sharing in the wealth, you are not going to be very happy. Seems pretty simple.

The Census figures show pre-tax income, so they are incapable of illustrating the effects of the 2001 and 2003 Bush tax cuts. The Bush administration argues that after-tax per capita incomes have actually gone up. You can get these numbers from the Bureau of Economic Analysis (BEA), and here they are (Disposable Personal Income in chained dollars from Table 7.1):


Wow. These inflation-adjusted figures seem to show a relentlessly improving income picture during the Bush years, despite the 2001 recession. Why do the two charts tell such different stories?

The BEA numbers differ from the Census numbers in several ways:

1. The BEA numbers are after-tax numbers, whereas the Census numbers are pre-tax numbers

2. The BEA numbers are per capita -- or average -- income figures, whereas the Census numbers are median values

3. The BEA numbers include all forms of income and compensation (including health insurance premiums paid by your employer, capital gains, food stamps, etc.), whereas the census numbers leave out those sources of income and quite a few others as well.

Which of these differences accounts for the different stories told by the BEA and Census income charts? People appear to be convinced that the second difference -- average (or per capita) values vs. median values -- explains it. An average value will go up even if the rich are the only ones who are doing better, whereas the median can remain unchanged or even go down. And it's the median value that best indicates how those in the middle are doing.

Here's Matthew Yglesias making the point that income figures like those supplied by the BEA are uninformative because they reflect per capita (or average) income:

Sadly, per capita numbers don't really tell you anything about how "most" people are doing. But here on the White House Economic Statistic Briefing Room website we have a link to median household income data. Median household income in 2004 was $44,389 which is a lot by world standards. But in 2003 it was $44,482 which was more. In 2002 it was $44,546 which was even more. In 2001 it was $45,062 which was even more. In 2000 it was $46,058 which was even more. In 1999 it was $46,129 which was even more. In 1998 it was $45,003 which was less, but still higher than today's median. And if you go all the way back to 1997, it was $43,430 -- lower than it is today. That's the sort of thing that probably lies behind dour economic sentiments.

Here is Jonathan Chait making a similar point :

You can have huge gains among a few households at the top and no gains among everybody else and still have a briskly rising average.

And an article in the Economist makes the same point:

Average after-tax income per person, Mr Bush often points out, has risen by more than 8% on his watch, once inflation is taken into account. He is right, but his claim is misleading, since the median worker -- the one in the middle of the income range -- has done less well than the average, whose gains are pulled up by the big increases of those at the top.

It's audacious to assert that Bush is being misleading without even looking to see if big gains by the rich really do explain the difference, but it's the kind of claim that liberal readers just gobble up (because they are already so sure that it is true). I've seen this claim made over and over again.

But it's all wrong.

The average BEA income figure is not increasing because the rich are pulling it up. It is increasing for some other reason, and that is what is so intriguing. To appreciate the fact that it is not being pulled up by the rich, all you need to do is plot incomes for the wealthy over the years in question. You could start with the Census data that everyone loves to cite. The census report presents income figures not only for those in the middle but also for those in the top quintile. Why doesn't anyone ever look at those numbers? Sure, there is simply no doubt that incomes for the wealthy are just skyrocketing, which allows Bush to mislead the nation by focusing on average incomes, but it would not hurt to actually prove it by showing the relevant figures. Why pass up a glorious opportunity to spotlight this glaring inequity in our economy over the last few years so that we can all seethe with anger at Bush for lining the pockets of his Halliburton buddies while leaving the rest of us high and dry? Here is that chart that shows how the wealthy are doing according to the Census figures:


Oddly, the rich incomes look a lot like the median incomes. Like everyone else, their incomes dropped after reaching record highs in 2000 and 2001.

How is it possible that average income increased up even in 2002, 2003 and 2004 if it was not being pulled up by big increases for the rich? In other words, what explains the difference between the income stories told by the Census median income figures (which everyone focuses on) and the BEA per capita income figures (which the Bush administration focuses on)? Well, it must be explained by the other two reasons (reasons 1 and 3) listed above. The BEA numbers are after-tax numbers, which are the numbers you want (as I have noted in several recent posts). And the BEA numbers include all sources of income, which is also a good thing.

In my next post on the economy, I'll show how all of the available data sets converge to suggest that the per capita BEA numbers offer the most valid picture of income levels for the average person. And it makes sense that these would be the valid income figures because they fit with all of the other statistics on the state of the economy. As George Will says in today's column entitled "Prosperity Amid the Gloom:"

It is said that workers' compensation has been stagnant. But to tickle that bad news from the statistics you must treat "compensation" as a synonym for wages, and then ignore the effect of taxation on individuals' well-being.

In a nutshell, he is correctly saying that the BEA income chart, not the Census income chart, captures the true state of affairs. The evidence presented here (with more to come soon) suggests that he's right.

The Scientific Flaws of the Lancet Study

Perhaps the Lancet study, which claimed that 655,000 Iraqis have died as a result of the war, has already been so completely discredited that no further comment is required. However, until now, no one has actually identified a methodological flaw in the study, so my assumption has been that the authors let their unconscious biases get the better of them when selecting households for interviews (thereby compromising the random sampling process). Now, however, someone who is very experienced in using the cluster sampling technique has identified a devastating methodological flaw:

However, the key to the validity of cluster sampling is to use enough cluster points. In their 2006 report, "Mortality after the 2003 invasion of Iraq: a cross-sectional sample survey," the Johns Hopkins team says it used 47 cluster points for their sample of 1,849 interviews. This is astonishing: I wouldn't survey a junior high school, no less an entire country, using only 47 cluster points.

Neither would anyone else. For its 2004 survey of Iraq, the United Nations Development Program (UNDP) used 2,200 cluster points of 10 interviews each for a total sample of 21,688.

Hmmm. 2,200 cluster points vs. 47. I'm thinking you go to the considerable trouble of using 2,200 cluster points when your goal is to find the truth. Better to use 47 when you have a different goal in mind.

The use of only 47 cluster points was not the study's most egregious flaw, though. The author goes on to say:

With so few cluster points, it is highly unlikely the Johns Hopkins survey is representative of the population in Iraq. However, there is a definitive method of establishing if it is. Recording the gender, age, education and other demographic characteristics of the respondents allows a researcher to compare his survey results to a known demographic instrument, such as a census.

Dr. Roberts said that his team's surveyors did not ask demographic questions. I was so surprised to hear this that I emailed him later in the day to ask a second time if his team asked demographic questions and compared the results to the 1997 Iraqi census. Dr. Roberts replied that he had not even looked at the Iraqi census.

And so, while the gender and the age of the deceased were recorded in the 2006 Johns Hopkins study, nobody, according to Dr. Roberts, recorded demographic information for the living survey respondents. This would be the first survey I have looked at in my 15 years of looking that did not ask demographic questions of its respondents. But don't take my word for it--try using Google to find a survey that does not ask demographic questions.

Iraq is in bad enough shape that we don't need political extremists pushing their crazy theories in once reputable medical journals to feed the anti-Bush jihad. The truth of the situation in Iraq is sufficiently bad right now that people here at home have plenty of ammunition they can use to relentlessly savage their own president as he tries to wage a difficult war and bring democracy to the vast majority of Iraqis who want to make it work.

Senate Race Update

In a prior post, I showed that individual polls are so variable that you cannot really tell very much from them. That's why it's always much better to rely on the average of multiple polls, as they do at RealClearPolitics.com. The other reasonable approach to forecasting the future is to see how people are betting at Tradesports.com.

In my previous update on the critical Senate races (almost 2 weeks ago), I showed poll results for the 7 states of most interest:

Missouri
Montana
New Jersey
Ohio
Pennsylvania
Rhode Island
Tennessee

These are the states where a party switch has, at one time or another, seemed like a real possibility. I also showed results for Connecticut and Virginia because there is a lot of talk about those two races even though the outcomes do not seem to be in serious doubt. Here is the table I showed on October 5:


The states are listed in red or blue. Red means that the Senate seat is currently held by a Republican, whereas blue means that it is currently held by a Democrat. The table shows who was ahead at the time according to the average of multiple polls from RealClearPolitics and according to the betting at Tradesports.com. So, the first state (Missouri) is in red because it is currently held by a Republican. The question marks off to the left for Missouri indicate that I considered it to be a tossup on October 5. I probably should have listed Tennessee and New Jersey as tossups at the time as well. Here is the latest polling and trading information on these Senate races:


Montana, Ohio, Pennsylvania, and Rhode Island are all red states that both methods predict will become blue states (so I put a blue asterisk off to the left). That's a pickup of 4 Senate seats for the Democrats. New Jersey is a blue state that both methods now say will stay that way (so it gets a blue asterisk, too).

The other two critical races are tossups (Missouri and Tennessee). Just flipping a coin for both races, there is a 25% chance that both will stay Republican, a 25% chance that both will be picked up by the Democrats, and a 50% chance that one will stay Republican and the other will be picked up by a Democrat. The Democrats need 6 seats to regain control of the Senate, which means that they would need to win all four of the races that lean heavily their way (Montana, Ohio, Pennsylvania, and Rhode Island) as well as these two tossup races (Missouri and Tennessee). If we assume a 100% chance of picking up the 4 seats that lean heavily their way and a 25% chance of also picking up both tossup seats, then it means that the Democrats have a 25% chance of picking up the 6 seats they need to control the Senate. In other words, the Republicans have a 75% chance of retaining control of the Senate. At Tradesports.com, the betting on the aggregate question of whether or not the Republicans will retain control of the Senate is at 65%, a bit below what I currently predict (the betting on the House is that there is only a 35% chance of Republicans holding on to power).

In the last presidential election, I learned that averaged polls have a great deal of validity (just as they theoretically should). Also, it has long been known that trading markets have a great deal of validity. However, averaged polls and trading markets are predictive just before the election is held. Are they predictive even 3 weeks out? That I don't know, but I am looking forward to finding out. Right now, the betting is that the Republicans will hang on to the Senate and lose the House.

A few days ago, Powerline asked if all of these polls can be screwy (and therefore not predictive of what will happen in November). Yesterday, they linked to a disussion about that by Dafydd ab Hugh, who argues that they might very well be screwy. He points out that pollsters weight their raw data based on predicted voter turnout:

Thus, the pollster must adjust the expected turnout model to take these changes into account; and this is where the bias creeps in, probably unbeknownst to the pollster: how much less turnout should we expect from evangelical voters in 2006 vice 2002? How much more turnout of women, or blacks, or Hispanics?

Similary, at TKS, Jim Geraghty thinks that the polls are problematic:

As I've mentioned from time to time, I have a theory that the traditional system of telephone polling is broken. I suspect that caller ID, cell-phone only homes, the general public’s busier lifestyle (who has time to spend a half hour answering a pollster’s questions?), suspicion of pollsters and the media, the spread of push-polling and a host of other factors have come together to make it nearly impossible to get a good representative sample.

All of this reminds me of very similar talk before the last presidential election. The polls were all over the place, and it seemed likely to me that cells phones and weighting methods were generating invalid results. That's why I was so interested to see how it would it would all turn out in the end. And, in the end, the averaged polls were spot on. It was truly amazing. All the talk of cell phones and weighting methods and the young turning out in unexpectedly high proportion -- none of that mattered. The polls (once they were averaged to reduce the amount of natural variability associated with a poll of only 1000 people) were just stunningly accurate. That's why I average polls today, and it's why I suspect that all of this talk about polls being invalid might be off the mark. The polls may not yet be predictive this far out from an election, but I believe they provide an accurate picture of where voters stand as of today.

There is good reason to believe that things will change in favor of the Republicans by election day because they have more money to wage a fight in the last few weeks of the campaign. This story from the tossup state of Missouri suggests that the race may not be so close by the time November rolls around:

Talent maintains wide lead in Senate money race
SAM HANANEL
Associated Press

WASHINGTON - Democratic Senate candidate Claire McCaskill outraised Republican Sen. Jim Talent by more than $1 million over the last three months, but Talent entered the campaign's home stretch with a huge financial advantage, the latest finance reports show.
...
But Talent's campaign had a whopping $4.2 million in cash to spend in the final weeks leading up to the Nov. 7 election, while McCaskill had $246,000 in the bank.

I am looking forward to seeing how it all turns out. For the time being, though, I think that you should believe what the polls are saying.

Even the (Serious) Left Dismisses the Lancet Report

Glenn Reynolds linked to a report at Iraq Body Count that is the most definitive refutation of the recent Lancet study that I have seen yet. The Lancet study claimed that 655,000 Iraqis have died as a result of the invasion, a preposterously high figure that no objective analyst could possibly take seriously. Iraq Body Count is a left-leaning, anti-Bush site, but they are not unserious liberals suffering from a bad case of Bush Derangement Syndrome (unlike the politically motivated authors of that study in Lancet). Instead, they are determined to track deaths in Iraq as objectively as they can, and I, for one, have long valued their efforts (just as I value the efforts of those at Iraq Casualty Count).

I was happy to see that the people at Iraq Body Count identified some of the same problems with the Lancet study that I did. For example, in my critique of the study, I said:

In other words, an astounding 92% of the deaths they documented were confirmed by a medical death certificate...But if they have official certificates for those deaths (which amazes me given the chaotic conditions there), it means that the deaths are being documented by medical officials. This leads to a natural question: What does the Iraq Ministry of Health say? They say this:

BAGHDAD, Iraq -- More than 2,660 Iraqi civilians were killed in Baghdad in September, according to new Health Ministry figures -- about 400 more than in the month before, despite an intensified U.S.-Iraqi sweep aimed at reining in violence.

But the Lancet study implies that, on top of these reported deaths, there were 27,000 more deaths in Iraq in September, 92% of which were documented by medical officials even though no one in the Health Ministry knows about it!

Similarly, Iraq Body Count had real problems with those death certificates:

Half a million death certificates were received by families which were never officially recorded as having been issued.

In 87% of cases where deaths were reported, the survey team asked to see death certificates, leading to the Lancet authors' statement that "92% of households had death certificates for deaths they reported". Assuming, as the authors do, that this is representative of the population as a whole, would imply that officials in Iraq have issued approximately 550,000 death certificates for violent deaths (92% of 601,000). Yet in June 2006, the total figure of post-war violent deaths known to the Iraqi Ministry of Health (MoH), combined with the Baghdad morgue, was approximately 50,000.

If the Lancet estimate is correct then it follows that either (a) 500,000 documented violent deaths, for which certificates were issued, have somehow managed to completely disappear without a trace to Iraqi officials or the international media or (b) there is a vast, elaborate, and very successful, cover up of this massive number of bodies and their associated paper trail being carried out in Iraq.

The Lancet study is clearly asking us to believe too much.

One wonders how that study could have erred so badly when they extrapolated from the 547 deaths they documented to the 655,000 deaths they estimated. About that, I said this:

If you read the Lancet report, you'll see that they actually documented 547 deaths in the post-invasion period and then extrapolated that number to nearly 655,000. That might be OK if their sampling process were truly random, but the authors themselves give one very good reason to believe that it wasn't. As they say on page 7:

Although interviewers used a robust process for identifying clusters, the potential exists for interviewers to be drawn to especially affected houses through conscious or unconscious processes.

This is why it is not really helpful to have hyper-partisans with a clear left-wing agenda conducting a study with an outcome that was obviously pre-ordained. Even relatively small biases along those lines would translate into huge increases in the number of estimated deaths. And it is very hard to imagine that these agenda-driven, left-wing authors (and their intrepid team of well-meaning researchers) did not suffer from biases that led to inflated casualty claims.

Similarly, Iraq Body Count said:

We would hope that, before accepting such extreme notions, serious consideration is given to the possibility that the population estimates derived from the Lancet study are flawed. The most likely source of such a flaw is some bias in the sampling methodology such that violent deaths were vastly over-represented in the sample. The precise potential nature of such bias is not clear at this point (it could, for example, involve problems in the application of a statistical method originally designed for studying the spread of disease in a population to direct and ongoing violence-related phenomena). But to dismiss the possibility of such bias out of hand is surely both irresponsible and unwise.

Note that this is not some rightwing web site shilling for George Bush. Here is one more paragraph from their press release that clearly shows where their political sympathies lie:

Do the American people need to believe that 600,000 Iraqis have been killed before they can turn to their leaders and say "enough is enough"? The number of certain civilian deaths that has been documented to a basic standard of corroboration by "passive surveillance methods" surely already provides all the necessary evidence to deem this invasion and occupation an utter failure at all levels.

Sounds like liberal boilerplate to me. But one wonders what they mean by "enough is enough." Do they mean "end the occupation now so that Shiites and Sunnis will stop killing each other"? This otherwise serious web site starts acting like a typical unserious Democratic candidate for congress here by complaining about what Bush has done without offering anything in the way of a useful strategy. Since a withdrawal of troops would, according to all accounts, lead to a drastic increase the number of deaths in Iraq, one almost gets the impression that Iraq Body Count is less concerned with the people who are dying in Iraq than they are with establishing once and for all that Bush was wrong to invade Iraq. I don't really believe that, but it is how they are behaving with their emotional, knee-jerk reaction ("enough is enough") that seems tantamount to saying "get out and let the unrestrained slaughter of civilians begin."

Despite these quibbles, the folks at Iraq Body Count have done a great job of refuting that ridiculous study in Lancet.

It's Bad in Baghdad

Halfway through the month of October, Iraq Casualty Count listed 761 civilians killed, which forecasts about 1500 killed by the end of the month. That would be the highest total yet recorded before morgue deaths are added in, and it suggests that sectarian violence has increased above the already high level of violence that has been in effect since the bombing of the Golden Mosque more than 6 months ago. Based on my reading of the casualty reports, the situation is not really best described as back-and-forth sectarian violence. Instead, it is better described as Shiite militias slaughtering Sunnis. If Baghdad is having a civil war, the Sunnis appear to be losing badly. It was always clear that it would be that way if the Shiites and Sunnis started fighting because the Sunnis are heavily outnumbered.

Of the 761 killed so far in the month of October, 364 appear to have died at the hands of Shiite militias. The stories associated with those killings are all like this:

BAGHDAD - At least 25 bodies were found in different parts of Baghdad over the last 24 hours, police said. The bodies bore signs of torture and had gunshot wounds.

That's the method used by Muqtada al Sadr's Mahdi Army. By contrast, only about 60 of the deaths seem to have been caused by Sunnis targeting Shiites, which we see in stories like this (I think):

BAGHDAD - A bomb placed under a car demolished a bakery in Baghdad's southern Doura district, killing 10 and wounding four, an Interior Ministry source said.

That's the method the Sunnis generally rely on. Although it is impossible to tell for sure, all of the other deaths appear to be criminal acts of violence or collateral casualties from roadside bombings of police or army patrols. There are also some mortar attacks in there as well that are hard to classify.

In any case, in Baghdad, the Shiites appear to be exterminating the Sunnis, and the Sunnis who are not being killed are being driven from their homes:

BAGHDAD, Iraq. Sunni Muslims were fleeing across the Tigris River on Monday, trying to escape a four-day rampage of sectarian fighting in their Shiite-dominated home city north of Baghdad. At least 91 people have died --— all but 17 of them Sunnis.

Sunnis, a minority in the city of Balad, said militiamen had been going door to door, giving them two hours to clear out of their homes, and one police officer said the bodies of the city's Sunni minority lay unclaimed in the streets.

I tend to believe the Sunni lawmakers who complain that Prime Minister Maliki -- himself a Shiite -- is turning a blind eye to the violence. For example, this story suggests that Miliki is playing a waiting game:

US commanders now cite Shiite militias as the biggest single threat to the stability of Iraq and warn that they are awaiting the green light from Prime Minister Nuri al-Maliki to launch an operation to clear Sadr City of gunmen.

Maliki has refused to give the green light for any operation into Sadr City for fear of destroying the whole neighborhood.

Is that really the reason why he is not giving the green light? It doesn't sound very compelling to me. Although Miliki is holding off, preparations are underway for something big. Thousands of US troops and thousands more Iraqi troops have moved into Baghdad. And they've been busy:

Pace said a berm designed to encircle Baghdad and restrict the movement of death-squad members and insurgents in and out of the city had been completed recently and that 28 checkpoints manned by Iraqis now controlled the entrances to the city.

Could it be that the berm is designed to entrap the militias inside Baghdad when Maliki finally does give the green light for a Sadr City crackdown? And is Maliki waiting to give that green light until the Sunni insurgency gives up its insane attempt to regain control of Iraq?

Both seem possible to me. Although the chaos in Iraq is dispiriting, it does not appear to be a serious threat to the democratically elected government so long as we stay, and I see no signs that Bush is about withdraw our troops. And the fact that Miliki is allowing the violence in Baghdad to happen (or so it seems to me) suggests that there may be a strategic purpose behind his unwillingness to finally approve a crackdown.

Just guessing here, but my theory is that the Sunnis are going to be slaughtered by Shiite militias until the Sunni insurgency gives up the fight. Maliki is, I assume, negotiating with the Sunnis and is giving them a clear choice: if you hold onto your dreams of returning to power, then you will watch your people be systematically annihilated by Shiite militias. But if you give up the fight, the government (backed by US forces) will finally crack down on those Shiite militias in Baghdad. Everything is in place for that crackdown to happen -- everything but a Sunni pledge to disarm and end the insurgency.

This is, of course, a speculative scenario pieced together from the various news reports I am reading. But if this theory is right, the Sunni insurgents hold the fate of the larger Sunni population in their hands. I hope they give up the fight, but I doubt they will until many more of their own people are summarily executed by Muqtada al Sadr's death squads.

Income Inequality was Much Greater Under Clinton Than Bush

In reading through internet discussion prompted by my last post, I see that some are referring to a June 2006 article in the Economist that ostensibly refutes my claim that the rich were better off under Clinton than Bush. I had come across that article independently a few days ago while searching for reasons why people believe that the Bush tax cuts provided a windfall for the wealthy.

If you believe that the Bush tax cuts have benefited only the rich, I sincerely hope that you will read the article in the Economist entitled "Inequality in America: The rich, the poor and the growing gap between them." If you read it carefully, you'll see that it is an indictment of Clinton, not Bush. Well, it is not rally an indictment of anyone, but if you think like many liberals do ("if the rich are making off with more of the money, the thing to do is to savage your president"), then it's time for you to trash Clinton for providing a windfall for the wealthy. I kid you not. The main evidence, which I present towards the end of this post, should make your liberal blood boil.

The Economist article begins by buying into a fallacy that I called the Wages and Productivity Hoax. I'm amazed to see such a basic mistake being made in a reputable publication like this, but there it is. In a nutshell, the flawed argument is that while wages and productivity were supposedly tightly connected in the 1990's, things mysteriously changed after 2000. Productivity continued to rise after 2000, but wages did not increase accordingly. As the article says:

But after 2000 something changed. The pace of productivity growth has been rising again, but now it seems to be lifting fewer boats. After you adjust for inflation, the wages of the typical American worker—the one at the very middle of the income distribution—have risen less than 1% since 2000.

There are multiple fast ones evident in this paragraph. First, it discusses wages (a pre-tax measure that ignores all non-wage income) when what matters to people -- obviously -- is after-tax income (from all sources). But it's a good trick to use when your goal is to mislead unsuspecting readers. Second, it overlooks the fact that wages and productivity have not been strongly correlated for many years, even for people with a college education. They were briefly correlated only during the dot-com bubble, which is why the article misleadingly uses that period for its baseline comparison. Everyone agrees that the bubble years (starting about 1997 and ending about 2001 or 2002) were not economically normal years. So, instead of focusing in on those years for comparison purposes (as the Economist does), let's zoom out a bit and look at the relationship over a longer period of time. Here is a graph from my prior post that illustrates this point with respect to wages for those with a college degree:


These values are normalized to 100 for year 1995 (productivity data found here and wages data found here). As you can plainly see, productivity has almost doubled since 1973, but college wages have ticked up only a notch, and even that uptick occurred mainly during the dot-com madness.

But if your goal is to mislead, it's best to ignore that fact and to pretend that what has happened since Bush took office is some kind of aberration. In fact, it was the late Clinton years that were aberrant, but for some reason, the Economist does not want you to know that (or perhaps they don't even know it).

The article then moves on to the next standard trick:

The White House professes to be untroubled. Average after-tax income per person, Mr Bush often points out, has risen by more than 8% on his watch, once inflation is taken into account. He is right, but his claim is misleading, since the median worker—the one in the middle of the income range—has done less well than the average, whose gains are pulled up by the big increases of those at the top.

This is false. The editors use pre-tax median income to make their claim, and they do not realize that those big gains at the top simply did not happen, as I explained yesterday (at least not as of 2003). You can analyze the data yourself to see that the rich made quite a bit less during the first 3 years of the Bush administration (though they were making off like bandits under Clinton), whereas the ones in the middle income range fared better. Again, it's a basic error to compare median pre-tax income with mean after-tax income. It makes a lot more sense to look at median after-tax income if you have a problem with the mean after-tax income figure cited by the White House. When you do that, you see that those in the middle are doing OK. It's the rich and the poor (not the larger group in the middle) who did not seem to benefit from the 2001 Bush tax cuts. I agree that the White House should avoid using mean income figures, and I do not really know why the mean figure shows such an increase, but I do know that it's not because incomes for the wealthy were skyrocketing. Those incomes came down after the 2001 tax cuts.

Finally, the article notes this:

The one truly continuous trend over the past 25 years has been towards greater concentration of income at the very top. The scale of this shift is not visible from most popular measures of income or wages, as they do not break the distribution down finely enough. But several recent studies have dissected tax records to investigate what goes on at the very top.
The figures are startling. According to Emmanuel Saez of the University of California, Berkeley, and Thomas Piketty of the Ecole Normale Supérieure in Paris, the share of aggregate income going to the highest-earning 1% of Americans has doubled from 8% in 1980 to over 16% in 2004.

Before you convict Bush of enriching the wealthy, take a look at the relevant graph (which I made using the data found here and then plotted for both pre-tax and after-tax income):


The trend is unmistakable. From the day Clinton took office in 1993, the top 1% started making off with an increasing share of the wealth. Using pre-tax income, which is what the Economist article apparently refers to, you can see that the top 1% made 12% of the money in 1993, but that steadily increased to record levels (to nearly 18%) in year 2000 -- when Bill Clinton was still president. But then it dropped when Bush took office.

Are you astonished by this? You should be. The 2001 Bush tax cuts were supposedly a windfall for the wealthy. Everyone seems to believe it, and I am truly shocked to learn that it is demonstrably false whether you look at income figures (as I did yesterday) or at the "share of wealth" figures. If the Bush tax cuts had been a windfall for the rich, don't you think that their incomes would have increased and that their share of the income pie would have increased as well? Both decreased after the 2001 tax cuts instead.

There is an amazing myth propagating through the mainstream media according to which the Bush tax cuts have benefited only the wealthy. As it turns out, income inequality skyrocketed to record levels throughout the Clinton administration and then decreased when Bush took office.

Obviously, the stock market has a lot to do with this. The rich are more invested in stocks than the rest of us, so when the market goes up, I assume that the rich make off with a bigger share of the overall pie. I, myself, am not savaging Clinton for this surprising statistical evidence that the rich made off like bandits during his administration. But I am saying that there has been no windfall for the wealthy on Bush's watch (i.e., his tax cuts did not even make up for their losses due to the stock market crash), and that it's time for you to start savaging Clinton if you have been mistakenly savaging Bush because of a nonexistent windfall for the wealthy that occurred due to his tax cuts. But don't worry, the stock market is on the rise again, and I suspect that the rich are doing quite well now because of it, so you'll get your chance to savage Bush, too. It's just that you've jumped the gun a bit, and you forgot to savage Bill Clinton while you were at it.

The Rich Were Better Off Under Clinton Than Bush

As I noted yesterday, a standard liberal talking point (one that is constantly reinforced by a sympathetic and uncritical media) goes like this:

Democratic leaders, candidates, and their surrogates are likely to maintain their argument that the Bush tax cuts in 2001 and 2003 resulted in a windfall for the wealthy and contributed to the relative stagnation in income for middle- and low-income families. They cite data showing that most American families saw their take-home pay lag behind inflation in the past five and half years, and they say that Republican tax cuts have made the distribution of wealth and income more unequal today than at any other time since the early 20th century.

This quote is taken from a new article by John Maggs in the National Journal, but I have seen the same point made in countless other articles and news stories.

Today, in an effort to investigate this claim in more detail, I look at after-tax incomes for the last 3 years of the Clinton administration (1998-2000) vs. the first 3 years of the Bush administration (2001-2003). As before, the data were taken from the Congressional Budget Office and were compiled by the Tax Policy Center. 2003 is the most recent year for which figures are available, and you can analyze the data yourself by going here.

In the first graph, I show the after-tax figures for each quintile of the population (where quintile 1 = the poorest 20%, and quintile 5 = the richest 20%). What we should see, if the Democrats and the media are to be believed, is that incomes were equitably distributed during the Clinton years and then became seriously skewed in favor of the rich (due to the 2001 tax cuts) during the Bush years. Here is what the data actually suggest:



The evidence simply does not support the claim. Can you imagine what it would be like if every article suggesting that the Bush tax cuts favored the wealthy included a chart like this? Well, the article would not work, so it will never happen.

As you can see, inflation-adjusted after-tax incomes for the middle quintiles (quintiles 2, 3 and 4, which make up the middle 60% of the population) remained at their historically high levels or even edged up a bit during the first 3 years of the Bush administration, this despite the 2001 recession. This completely contrasts with the widespread belief that average folks have been falling behind as prices for healthcare and education have increased. Even more astonishing to me is the fact that incomes for the richest among us (quintile 5) have actually dropped during the early Bush years -- the exact opposite of what liberal gospel would have you believe. To put this another way (and this is the main point of today's post), the richest 20% of the population fared better under Clinton than they did under Bush. The poorest 20% did, too, but as I explained yesterday, most of their losses occurred during Clinton's last year in office (losses that continued during the Bush years). Still, if my analysis is accurate, it would be fair to say that by 2003, the Bush tax cuts had not helped to maintain the incomes of the poorest 20%.

My main point is that the Democrats' claim that the Bush tax cuts helped only the wealthy does not square with the facts, and I am both astounded to discover that this is true and discouraged to know that there is no possible way the American public will ever appreciate it. If the media hasn't figured it out by now, they never will. Instead, they will continue to focus on wages (i.e., on only part of the income story) or on pre-tax incomes (as if that could tell you about the effects of the tax cut) to tell a story that just makes perfect sense to the liberal mind. Liberals believe that they have George Bush pegged, and one of things they just know about him is that he doesn't care about the little guy because he's too busy lining the pockets of his rich Halliburton buddies. It's perfectly natural to fall prey to confirmation biases (i.e., to uncritically accept evidence that supports your preconceived notions while filtering out contradictory evidence), and I believe that explains why the media keeps bringing you a story that appears not to be true. The existence of confirmation bias is the main reason why it would be better to have a mainstream media composed of both liberals and conservatives.

The analysis presented above focused on the wealthiest 20% of the population. Perhaps you think that it is not the wealthiest 20% that made off like bandits due to the 2001 Bush tax cuts but that it was the wealthiest 10% (or the wealthiest 5% or the wealthiest 1%) instead. If that is what you think, my question for you is this:

Why do you think that?

I am sure it would not be hard to find a claim to that effect made by a mainstream media reporter, but that doesn't mean anything. What evidence convinces you that it is true? Here is the evidence that convinces me that it isn't true:


Even the wealthiest 1% (especially the wealthiest 1%) were better off under Clinton than under Bush. Yet the media would have you believe that the Bush economy -- unlike the egalitarian Clinton economy -- is further enriching the wealthy at the expense of everyone else. And they would have you believe it because it's what they, themselves, sincerely believe to be true. As best I can determine, what they believe to be true (and what they are trying to persuade the American public is true) is actually false.

Perhaps the 2003 tax cuts (unlike the 2001 tax cuts) enriched the wealthy while leaving the rest of us behind, but the data are not available to say one way or the other (at least not that I can find). I can hardly wait to find out the truth of the matter. Unfortunately, if the past is any guide, the facts will be available only after the media regales us with hundreds of stories that knowingly explain how Bush's tax cuts did little more than provide "a windfall for the wealthy."

Judging the Republican Economy

John Maggs has an excellent new article in the National Journal entitled "What If It's The Economy?" His article is also part of a series of headline stories this morning at MSNBC.com about The Republican Economy. In his article, he says:

Democrats say they would welcome an election that turns on the economy, and they make their own case that most Americans have seen their living standards erode during Republican rule, largely because of tax cuts that the Democrats claim were skewed toward the rich.

That claim is my focus today because, as best I can tell, it is not actually supported by any factual information. The article goes on to say:

On the day of the Ohio visit, Democratic National Committee Chairman Howard Dean put out a statement titled "Americans Not Buying Bush's PR Campaign on the Economy," and used Census Bureau data to argue that household income dropped for most workers last year.

As I said before, it is downright bizarre for Democrats to attack the Bush tax cuts as helping only the wealthy by citing pre-tax Census income figures. Obviously, the after-tax figures for most workers are what you want to know if your goal to assess the effects of the Bush tax cuts.

Here's more:

Democratic leaders, candidates, and their surrogates are likely to maintain their argument that the Bush tax cuts in 2001 and 2003 resulted in a windfall for the wealthy and contributed to the relative stagnation in income for middle- and low-income families. They cite data showing that most American families saw their take-home pay lag behind inflation in the past five and half years, and they say that Republican tax cuts have made the distribution of wealth and income more unequal today than at any other time since the early 20th century.

Does anyone ever stop to ask whether the tax cuts actually did benefit only the wealthy by looking at pre-tax and after-tax income for the various income groups? Or do they just take it for granted that it happened based on what they know to be true about George Bush? I think it might be the latter.

Yesterday, I showed the income data for the middle quintile (those falling in the 40% to 60% range) based on data found here. The data seem trustworthy because they are from the Congressional Budget Office and were compiled by the Tax Policy Center, which is a joint venture of the Urban Institute and the Brookings Institution. They describe themselves like this: "The Center is comprised of nationally recognized experts in tax, budget, and social policy who have served at the highest levels of government."

My previous analysis of their data showed that while median pre-tax income dropped in the last few years, after-tax income remained more or less constant (at historically high levels). This graph shows the same thing, but for the 3 middle quintiles combined (i.e., for people in the 20% to 80% range -- excluding the poorest 20% and the richest 20%):


The red symbols indicate recession years. As you can see, for this large segment of society, after-tax income has held steady near their historically high levels in the last few years even though pre-tax incomes fell during that period. If you want to make it sound negative, you can say that incomes are "stagnant" because that creates the impression that incomes are going nowhere while inflation marches on (hence the ostensible "erosion" of living standards). But, according to these numbers, the reality is that after-tax incomes have remained at their highest levels ever and are staying there even after adjusting for inflation. This would appear to suggest that the Bush tax cuts helped average households, and it argues strongly against the idea that "most Americans have seen their living standards erode during Republican rule." It only looks that way if you look at pre-tax income (which is an odd thing to do when complaining about the Bush tax cuts). But that's what the Democrats and many others do because (I assume) they realize that most Americans won't detect their deceptive maneuver.

The next graph shows the same data for the poorest 20%:


Here, we can see that their after-tax incomes did not hold steady. Instead, their incomes have clearly fallen over the last few years. The Bush tax cuts did not help the poor directly, though they don't pay much in the way of taxes, so perhaps that should not come as a great surprise. Presumably, their lot will improve as the economy rebounds (though whether or not that happened in 2004, 2005 and 2006 remains to be seen because we don't have the numbers yet). Still, it seems perfectly accurate to say that the purchasing power of the poor plummeted during Clinton's last year in office (2000) and that the Bush tax cuts did not stop that decline following the 2001 recession.

The next graph shows how the wealthiest 20% made out, and this is what I really want to show you today. Surely the Bush tax cuts caused their after-tax incomes to skyrocket in the last few years. They are ones who making off like bandits. Right? That's what the Democrats want you to believe, but I never see them referring to any actual evidence. What are they basing their claims on? When you look at the evidence, as I do now, you discover that the after-tax incomes of the wealthy decreased after the Bush tax cuts, just as with the poorest 20%:


I'll have more to say about this soon, but, for the moment, the point is that the data simply do not support the idea that the Bush tax cuts have been a windfall for the rich. Instead, it seems more accurate to say that, by the year 2003, neither the rich nor the poor showed a direct benefit from the 2001 tax cuts, but the larger group in the middle did.

As best I can tell, there is no secret income for the wealthy that is being overlooked by this analysis. Here is what the incomes in question consist of:

Pretax cash income is the sum of wages, salaries, self-employment income, rents, taxable and nontaxable interest, dividends, realized capital gains, cash transfer payments, and retirement benefits plus taxes paid by businesses (corporate income taxes and the employer's share of Social Security, Medicare, and federal unemployment insurance payroll taxes) and employees' contributions to 401(k) retirement plans. Other sources of income include all in-kind benefits (Medicare, Medicaid, employer-paid health insurance premiums, food stamps, school lunches and breakfasts, housing assistance, and energy assistance).

The 2001 tax cuts did not differentially benefit the wealthy. Instead, they differentially benefited the large group in the middle. I find this result to be simply amazing in light of the widespread conviction on the left that the Bush tax cuts are helping only the wealthy, leaving the lower 80% to suffer from an eroding standard of living. Unless there is something wrong with the numbers compiled by the Tax Policy Center, it just does not appear to be so.

I'm Sad About Air America

As has been rumored for a while now, Air America filed for Chapter 11 bankruptcy protection. It's fate always seemed pre-ordained to me because there was never any reason to believe it would succeed. Since we are already saturated with left wing analysis of the news (provided non-stop by the mainstream media), there was simply no unoccupied niche to fill. By contrast, the virtual absence of right wing reporters left a giant niche that has been occupied by the likes of Rush Limbaugh, Sean Hannity, Hugh Hewitt, Dennis Prager, Michael Medved and many others.

But I'm sad to see Air America going down the tubes. I listen to them every day, and I hope they find a way to continue.

Once upon a time, being liberal meant, in part, being more concerned about the welfare of the disadvantaged (e.g., the forgotten poor, under-powered women, oppressed blacks, etc.) than conservatives typically are. Those are still positive qualities associated with liberalism in my book, and you can sometimes hear liberals on Air America discussing issues along those very lines (e.g., the fact that many people of lesser means do not have health insurance and what should be done about it). I feel my liberal past tugging at me when I listen to discussions like that. At other times, you can hear thoughtful liberal arguments against the invasion of Iraq and against taking a tough stand against Iran. I love listening to such discussions even if I do not happen to agree with them.

But it never takes long for the talk to turn to what liberalism has become today. Being liberal today is to harbor a deep conviction that Bush misled this nation into war, that Bush just wants to line the pockets of his Halliburton cronies, that Bush is incompetent, that Bush hates blacks, that Bush's economic policies are intentionally designed to favor the wealthy at the expense of the poor, that Bush wants to torture detainees, that Bush wants to trample your civil rights, and on and on. Randi Rhodes is amazing in this regard (but I admit she can be funny when she talks about the "Torturer in Chief").

I used to think that liberals were inclined to think and reason based on scientific evidence and logic, whereas conservatives were more likely to allow their thinking to be governed by religious convictions. Now, though, liberals have become mostly unconcerned about evidence-based reasoning because they are so incredibly prone to uncritically accept anything that places George Bush in a negative light. That's why they are capable of completely disregarding -- to my once utter disbelief -- the bipartisan Senate Intelligence Committee Investigation that dispensed with the "Bush lied!" nonsense and (essentially) convicted Joseph Wilson of lying about it. When I saw that report having no influence whatsoever on the thinking of liberals, I realized that, when it comes to George Bush, many of them simply cannot be reasoned with. They stuck with the Joe Wilson fantasy to the bitter end, and I would not be surprised to find that many still believe that Wilson proved that Bush misled this nation into war.

To be liberal today is to harbor a deep conviction that Bush is the root of all evil. And that's the liberalism that left me behind.

Although I very strongly side with the Republicans on issues of national defense, I'm still liberal in many ways: I don't favor outlawing abortion at conception, and I never will. I do favor programs aimed at helping black America in recognition of the fact that 200 years of oppression cannot be easily undone, and we haven't undone it yet (diversity-based affirmative action is another matter, and I do not favor that at all). Regulating assault rifles does not seem like an insane course of action to me. I also favor raising the minimum wage a bit, even though it seems clear that it does not help the poor (but it doesn't hurt them either so long as you don't raise it too much, so the gesture of concern would be worth it). My stand on the death penalty is influenced by the scientific evidence pertaining to its deterrence effect. And I would vote in favor of gay marriage even though I admit that the arguments against it seem pretty compelling to me, at times.

Despite those positions, few today would consider me liberal because (a) I don't hate Bush and (b) I believe that Bush is a principled individual who says what he means. Because I think along those lines (and because my position on national defense is like that of George Bush and Hillary Clinton), my liberal colleagues regard me as a radical conservative.

But I'm not really so conservative, and that's why I like listening to Air America (at least when they are talking about something other than what a monster George Bush is). The fact that I'm going to miss Al Franken and Randi Rhodes if Air America goes under must say something about my liberal bona fides.

Household Income and Economic Pessimism

I remain fascinated by the opinion of many on the left that our humming economy is benefiting only the wealthy. The crux of the issue is household income, a topic that I have considered several times and consider again today.

It could be true that the wealthy are making off like bandits while everyone else is left holding the bag, but I can't figure out how anyone could have a strong opinion on the matter. It is easy to find stories in the media that make unsubstantiated suggestions along those lines, but it is really hard to find the actual evidence you need to form an opinion one way or the other.

Here, again, is a figure from the Census Bureau showing median income over the years (found on page 4 of this report):


You can see that median income peaked in 1999 and the dropped each year until 2005, when it finally turned around (slightly).

Unfortunately, this is a pre-tax measure, so it is unable to show the effects of the Bush tax cuts that were enacted in 2001 and 2003. Even so, many people use this figure to explain why people are so down on America's undeniably fabulous economy. For example, reasonably sensible left winger Matthew Yglesias, in a post entitled Things are Great!, said a while ago:

Median household income in 2004 was $44,389 which is a lot by world standards. But in 2003 it was $44,482 which was more. In 2002 it was $44,546 which was even more. In 2001 it was $45,062 which was even more. In 2000 it was $46,058 which was even more. In 1999 it was $46,129 which was even more. In 1998 it was $45,003 which was less, but still higher than today's median. And if you go all the way back to 1997, it was $43,430 -- lower than it is today.

That's the sort of thing that probably lies behind dour economic sentiments. Lots of people -- most, really -- haven't been doing all that well.

I am amazed at how many people are ready to uncritically accept these numbers as telling you all you need to know. How does he know, for example, that disposable income (i.e., what we have left after taxes) went down in those years? Disposable income is what really matters, and this chart does not give you any information about that. In light of the Bush tax cuts, median disposable income could be doing quite a bit better than this chart implies. If you are a typical liberal, you are sure that Bush's tax cuts are helping out only the wealthy (at least that's how it seems to me). Perhaps that's true, but wouldn't you want to see after-tax incomes for the median household going down each year to prove that you are right? Pre-tax information doesn't tell you what you need to know to prove that Bush is evil personified.

Although the best way to measure income turns out to be an incredibly complicated issue, I am increasingly convinced that the household income figures supplied by the Congressional Budget Office provide the best snapshot of the income situation. So far, though, the data are available through 2003 (found here). Income figures are provided for each quintile of the population, and I used the data to compute both pre-tax and after-tax income for the 40%-60% income group. Values for this "mid-quintile" group are essentially the same as median income values. Here is what I found:


The green line shows pre-tax data, the blue line shows after-tax data, and the red dots indicate recession years (and the numbers have been adjusted for inflation due to increasing healthcare costs and the like). As you can see, the pre-tax line looks a lot like the median income chart provided by the Census Bureau. It peaked in 1999, and came down every year thereafter. Incomes for 2002 and 2003 (the last 2 points) are quite a bit below the preceding few years. So, when the stock market bubble burst, incomes that had been artificially inflated by that phenomenon came down.

But the after-tax values tell a somewhat different story. They remained at the historically high level of about $45,000 (though incomes in 2001 were slightly higher), and, unless I am missing something, this could have only happened because of the Bush tax cuts. It is therefore ironic to see people lamenting the income figures that they believe have been declining because the Bush tax cuts have helped only the wealthy (a conclusion they arrived at by looking only at pre-tax income figures!). When you look at the numbers the right way, it becomes clear that disposable income actually stayed up near the record high levels of the bubble years, all because of the Bush tax cuts.

This analysis fits with this study by the Joint Economic Committee that I showed a few days ago:


This graph shows median income for families with two children. The income values are a bit higher because it does not include all households, but the story is exactly the same: due to the Bush tax cuts that supposedly helped only the wealthy, median family income remained at historically high levels throughout the slowdown of 2001 and 2002 (and edged up to a new high in 2003).

As best I can determine, and as I explained in detail here, median incomes in the US are well ahead of median incomes in other major economies of Europe:


All of these measures are relative to the U.S. median disposable income (which is set to 100). The disposable income values come from the Luxembourg Income Study, and the European numbers represent the aggregate values for the Netherlands, Sweden, Germany, Finland U.K., and Belgium (which are the European countries for which these measures are available). The middle two bars of the figure show that median income in the US is substantially higher than in other advanced economies of the world.

Is that a reason for pessimism? The highest median income in the world, the highest ever in our great nation (except for a bubble economy year of 2001), and stumbling only slightly after the 2001 recession? I suspect that the numbers look even better for 2004 and 2005, but only time will tell. If the data from those years show plunging after-tax incomes, I'll see some evidence for the widely held view that folks in the middle are struggling with economic hard times. For the moment, though, I remain of the opinion that Americans are down on their fabulous economy because of the ridiculous way it is portrayed by a well-meaning but misinformed left-leaning media.

Lancet: Wildly Estimating 30,000 Iraqis Killed Per Month

I am, as always, amazed by the reaction on the left to news that can be used to place George Bush in a negative light. The new study in Lancet claiming that the invasion of Iraq resulted in 655,000 extra deaths over the last 3 years has been met with instant, uncritical acceptance on the left.

Imagine that a hyper-partisan on the right with a history of conducting pseudo-studies reporting nothing but "happy talk" on Iraq had conducted a brand new study showing that, amazingly, there are only 200 extra deaths occurring in Iraq each month. Supporters of the war in Iraq would not just uncritically gobble this up. Instead, they would laugh it off in the same way that they are laughing off the new study in Lancet. The reported death rate would be regarded as preposterous because it is so far out of synch with all other methods of measuring casualty statistics (which estimate between 2000 and 3000 killed per month). The same holds true of the Lancet study.

If you read the Lancet report, you'll see that they actually documented 547 deaths in the post-invasion period and then extrapolated that number to nearly 655,000. That might be OK if their sampling process were truly random, but the authors themselves give one very good reason to believe that it wasn't. As they say on page 7:

Although interviewers used a robust process for identifying clusters, the potential exists for interviewers to be drawn to especially affected houses through conscious or unconscious processes.

This is why it is not really helpful to have hyper-partisans with a clear left-wing agenda conducting a study with an outcome that was obviously pre-ordained. Even relatively small biases along those lines would translate into huge increases in the number of estimated deaths. And it is very hard to imagine that these agenda-driven, left-wing authors (and their intrepid team of well-meaning researchers) did not suffer from biases that led to inflated casualty claims. In fact, their claims turn out to be even wilder than most people realize.

The 655,000 post-invasions deaths have ostensibly occurred over the last 3.5 years, which comes to an average of more then 15,000 deaths per month. Anyone who has been following casualty statistics in Iraq, as I have, knows that the civilian casualty rates have effectively doubled since the bombing of the Golden Mosque at the end of February. Here is a graph using the casualty statistics from Iraq Casualty Count that shows this sad phenomenon:


Why does this matter? Because if there have been 15,000 deaths occurring each month all along since the invasion in 2003, it must now be the case that there are more like 30,000 deaths occurring each month. In fact, if I am reading Table 3 of the Lancet study correctly, their own data imply that nearly 400,000 of the extra deaths occurred in the most recent 13 month period (June '05 to June '06 -- the study ended in July of '06). The excess mortality rate for period 1 (March '03 to April '04) was 3.2 per thousand. If we assume 28 million Iraqis (a little above the 27.1 million figure from 2004 shown in their Table 1), that comes to 89,600 extra deaths in that period. The excess mortality rate for period 2 (May '04 to May '05) was 5.6 per thousand, which comes to 156,800. And the excess mortality rate for the most recent period (June '05 to June '06) was 14.2 per thousand, which comes to 397,600. In all, this comes to 644,000 extra deaths, which is close to the 655,000 figure they report.

The key point to appreciate is this: the Lancet study claims that there were nearly 400,000 extra deaths from June '05 to June '06, which comes to about 30,000 deaths per month over that 13-month time period!

That's the current death rate in Iraq, according to this study. Mysteriously, though, these bodies are not showing up in the Baghdad morgue or at hospitals or anywhere else. They are not being documented in any way.

But wait! Maybe they are being documented after all. The Lancet study says:

At the conclusion of household interviews where deaths were reported, surveyors requested to see a copy of any death certificate and its presence was recorded.

And later they say this:

However, certificates might not be issued for young children, and in some places death certificates had stopped being issued; our 92% confirmation rate was therefore deemed to be reasonable.

In other words, an astounding 92% of the deaths they documented were confirmed by a medical death certificate. The authors needed these death certificates because, otherwise, their results could be dismissed on the grounds that people are recalling many more deaths than really occurred. But if they have official certificates for those deaths (which amazes me given the chaotic conditions there), it means that the deaths are being documented by medical officials. This leads to a natural question: What does the Iraq Ministry of Health say? They say this:

BAGHDAD, Iraq -- More than 2,660 Iraqi civilians were killed in Baghdad in September, according to new Health Ministry figures -- about 400 more than in the month before, despite an intensified U.S.-Iraqi sweep aimed at reining in violence.

And we have this from the story as well:

Hospitals reported 1,196 fatalities from explosions, shootings and other attacks in September, and the morgue listed 1,471 people who died from violence.

All of this fits with my own earlier analysis of the September in Iraq, which you can find here. But the Lancet study implies that, on top of these reported deaths, there were 27,000 more deaths in Iraq in September, 92% of which were documented by medical officials even though no one in the Health Ministry knows about it!

Either that, or some authors with a well-meaning but extremely left-leaning political agenda allowed their unconscious biases get the better of them and wildly missed the mark when they extrapolated from the 547 deaths they documented to the 655,000 deaths they reported.

Like Fox News, I report, you decide.

About that New Study in Lancet

You may have heard this startling news:

Reporting this week in the online edition of The Lancet, a leading British medical journal, the researchers estimated that 654,000 more Iraqis died of various causes after the invasion than would have died in a comparable period before.

The Lancet, of course, is the same journal that published a similarly flawed study on the eve of the last presidential election in a transparent attempt to influence the outcome (to no avail, fortunately). That study claimed that more than 100,000 Iraqis had died as a result of the invasion.

It's bad enough that a once reputable journal would publish methodologically flawed studies as if they were methodologically sound. Publishing politically motivated, methodologically flawed studies is even worse. And the fact that it would habitually do so just before elections are held in America in an effort to ensure that Democrats win is indicative of a deep pathology on the part of the journal editors. They'd be much better off publishing a medical inquiry into Bush Derangement Syndrome. Lancet is, after all, a medical journal.

The Lancet has an even longer history of publishing inflated casualty claims from Iraq. Back when Saddam was in power, liberal outlets (including the Lancet) were also doing body counts. In an article in the Nation entitled "A hard look at sanctions in Iraq," we have this:

The grim question of how many people have died in Iraq has sparked heated debate over the years. The controversy dates from 1995, when researchers with a Food and Agricultural Organization (FAO) study in Iraq wrote to The Lancet, the journal of the British Medical Society, asserting that sanctions were responsible for the deaths of 567,000 Iraqi children.

I think this old figure might explain the new study. In fact, I can just hear the conversation that preceded it:

Lancet: Team, someone just reminded me that we once said that sanctions killed 567,000 Iraqi children. If we don't act fast, someone might credit George Bush with saving that many children now that child mortality rates in Iraq have come way down.

"Researcher:" That would be horrible! Fear not, I'm on it. Soon, I'll have a new "study" for you that beats that old figure. In fact, I guarantee that it will come in well above 600,000.

["study" conducted here]

Lancet: Whew! We did it. Now we can say that "research proves that George Bush killed more Iraqis than he saved." Thank God we got this vital information before Americans go to the polls.

New Jersey and Connecticut

The New Jersey Senate race is interesting because it was long considered to be the one seat that might switch from Democratic to Republican hands. Thomas Kean held the lead for quite some time, but all that changed recently. Here is a nice chart of averaged poll results that I found at RealClearPolitics:



You can see that Kean was leading for a long time, but then something happened and the lead crossed over in the first week of this month. Two notable events occurred near the time of the crossover:

1. On September 27, it was reported that "U.S. Sen. Robert Menendez's closest political adviser was secretly recorded seven years ago boasting of political power and urging a Hudson county contractor to hire somone as a favor to Menendez".

2. On September 29, Mark Foley resigned because of those instant messages he sent to young pages.

In light of those events, New Jersey voters could either decide to vote for Kean because Menendez himself was (possibly) corrupt, or they could decide to vote for Menendez because Foley is a Republican and Kean, being a Republican himself, is guilty by association. It's weird, but they seem to have chosen the latter option.

Could the situation in Iraq be affecting the race? Possibly, but I doubt it. If that were a general factor, Ned Lamont might be expected to close the gap on Lieberman as time goes on, but here's another chart from RealClearPolitics that shows just the opposite:



I take this to mean that the voters of Connecticut had fun toying with the silly idea of electing a complete nobody to the Senate for no other reason than he opposes Bush on Iraq, but they are (I hope) getting more serious as the election approaches. I can only hope that the same phenomenon begins to affect the other Senate races, including the one in New Jersey, in the next 4 weeks. As I have said before, liberals are not serious these days. They mostly stand on the sidelines screaming "you're doing it wrong," whether it be the economy, Iraq, the war on terror, missile defense, Katrina, or the Dubai ports deal (still a favorite subject on the left).

How are they going to fix our fabulous economy? What, exactly, will they do about the situation in Iraq? Where do they stand on missile defense? What's the harshest interrogation technique of high-level al Qaeda detainees they would allow? Voters have no idea, so, if the polls are accurate, they will be voting based on little more than blind faith in Democratic candidates (and in people like Harry Reid and Nancy Pelosi).

My own feeling is America will get serious as the election approaches and decide not turn the Senate over to the constantly carping Democrats, but that's just a feeling. It's useful to have people standing around complaining about everything, but giving them actual power is another matter altogether. I'm not sure the American public is ready to do that. Bill Kristol made a related point here:

And, you know, they can. They have a lot of money. And the good thing about elections is, at some point, the mainstream media loses control and people are actually entitled to use the contributions they've raised to go up on the air with ads. And you can go up with ads contrasting your positions with that of your opponent on these core issues of judges, taxes and terror. And I think Republicans then hold the Senate.

That's what I think, too, but I also know how impossible it is to really predict the outcome with any degree of confidence this far out. Even so, I'll be watching the various indicators and posting updates as election day approaches.

The Muqtada al Sadr Maneuver

In a new column, Fareed Zakaria says:

It is time to call an end to the tests, the six-month trials, the waiting and watching, and to recognize that the Iraqi government has failed. It is also time to face the terrible reality that America's mission in Iraq has substantially failed.

Let's see, the new government took office on May 26, 2006, so Zakaria is basically saying: "You've had 4 months and 13 days to fix this problem -- time's up!"

That's ridiculous. He should read the recently released NIE report to see why a premature withdrawal is the worst possible course of action. It said:

Should jihadists leaving Iraq perceive themselves, and be perceived, to have failed, we judge fewer fighters will be inspired to carry on the fight.

So the Zakaria plan is to immediately grant the jihadists a perceived victory by cooperating with their demands that we surrender now and withdraw from Iraq, thereby inspiring more fighters to carry on the fight.

Excellent idea. As liberal David Ignatius observed in a recent column entitled The Big Question Democrats Are Ducking:

Here's a reality check for the Democrats: There is not a single government in the Middle East, with the possible exceptions of Iran and Syria, that favors a rapid U.S. pullout from Iraq. Why? The consensus in the region is that a retreat now would have disastrous consequences for America and its allies. Yet withdrawal is the Iraq strategy you hear from most congressional Democrats, whether they call it "strategic redeployment" or something else.

A reality check is not something the Democrats crave. Instead, they lust after any strategy that will ensure that Bush's "adventure in Iraq" will go down in history as a failure, regardless of the strategic consequences.

As Zakaria sees it, there can be no negotiated solution in Iraq because:

If you were a Sunni, having watched government-allied death squads kill and ethnic-cleanse your people, would you accept a piece of paper that said that this government will now give you one third of Iraq's oil revenues if you disarm?

No, but what Zakaria overlooks is that a share of oil revenues may not be all that Prime Minister Maliki is offering to the Sunnis. He may also be offering to put an end to Muqtada al Sadr's reign of terror -- but if and only if the Sunnis finally give up the fight.

Two recent news stories seem to be consistent that idea (an idea I first began to entertain here). First, this news story suggests that there is a looming militia crackdown in Baghdad, but one point made in the article seems to support my claim that we are reluctant to take out Muqtada al Sadr's militia because to do so would be to selectively take out only half the problem:

“I can’t drive (the Mahdi Army) into the dirt and let (al-Qaida) basically conduct suicide attacks at will,” one senior coalition intelligence officer said on condition of anonymity for security reasons. “I’ve got to take both elements out of the equation.”

That makes a lot of sense to me, though I suspect that there is a little more to the story. Specifically, the Iraqi government may have decided not to move on the Mahdi Army in a serious way until we have elicited some real concessions from the Sunnis -- concessions that seem more likely in light of changing attitudes that have occurred since Muqtada al Sadr began executing large numbers of Sunni civilians. As I noted before, polls show that the Sunnis are no longer quite as insistent that Americans troops leave the country, with the percentage calling for withdrawal within six months dropping from 83 percent a while ago to 57 percent more recently. That remarkable change in attitude may foreshadow an increased willingness on the part of the Sunni insurgency to rethink their stubborn (and truly hopeless) position. This LA Times article suggests that we may be seeing progress on that front:

Tribes Heed Call to Join Battle for Iraq
Maliki enlists the Sunni groups in an attempt to clear insurgents from Al Anbar province.
By Kim Murphy, Times Staff Writer
October 5, 2006

BAGHDAD — As tribal leaders from Iraq's troubled Al Anbar province met last week with Iraqi Prime Minister Nouri Maliki, pledging their support to clean out Al Qaeda insurgents, it soon became clear that they were as good as their word.

And why are they becoming cooperative? Because of the brutal execution campaign being waged by Muqtada al Sadr:

U.S. officials say the decision of some tribal leaders to begin going after insurgents reflects growing public anger over attacks that have killed or injured more than 8,000 Iraqis, according to local government figures.

I don't think that the Sunnis are angry about attacks by al Qaeda and Sunni insurgents that have killed Shiites. Instead, I assume they are angry that Muqtada al Sadr's Mahdi Army is retaliating by summarily executing large numbers of Sunnis. And they are not just angry; they are also afraid (hence their willingness to deal with Prime Minister Maliki).

As noted in a recent Defense Department report, the Mahdi Army and the Sunni extremists rely on different methods in their fight against each other. In particular, the Mahdi Army has started "...a campaign of overt executions and mass kidnappings of Sunni civilians" while the Sunni extremists rely on "...large scale and mass-casualty bombings of Shi'a gatherings" (p. 34). Muqtada al Sadr's execution campaign seems to have the upper hand, and Prime Minister Maliki's deal to the Sunnis appears to be this: give up the fight, and al Sadr's Mahdi Army will be disarmed and will no longer be allowed to do this:

01 Oct 2006
BAGHDAD - A total of 50 bodies were recovered by Baghdad police in various parts of the city over the 24 hours to Sunday evening, an Interior Ministry official said. Many had been tortured and most were bound and shot in the head.

02 Oct 2006
BAGHDAD - The Iraqi Islamic Party, the most prominent among the Sunni minority, said all 26 people kidnapped from a meat processing factory in Baghdad's Amel district on Sunday were Sunnis and were found dead in south Baghdad on Monday.

05 Oct 2006
BAGHDAD - A total of 30 bodies, most of them shot and tortured, were found in different districts of Baghdad during the past 24 hours, a source in the Interior Ministry said.

07 Oct 2006
BAGHDAD - Police said they had found 51 bodies, many bound and tortured, in the capital in the past 24 hours.

10 Oct 2006
BAGHDAD, Iraq - Authorities found the mutilated bodies of 60 men in Baghdad in the 24 hours ending Tuesday morning, likely the latest victims of the sectarian death squads that roam the capital.

That's how things are going for the Sunnis so far in the young month of October. By refusing to rein in Muqtada al Sadr, Prime Minister Maliki may be giving the Sunnis a glimpse of the grim future they face. It is, perhaps, his method of finally defeating the Sunni insurgency, and, for the time being, Muqtada al Sadr would appear to be an essential part of the equation.

Think Missile Defense this November

North Korea just successfully tested a nuclear bomb. They, like Iran, decided long ago that so long as the world takes a diplomatic approach to the problem, nothing will stop them from making their nuclear bomb. It's a tried and true strategy, one that will result in a nuclear-armed Iran in the not-too-distant future.

With regard to the national security of the Unites States, ballistic missile defense suddenly becomes an even higher priority. North Korea has its bomb, and they are trying hard to perfect a ballistic missile that can reach the west coast of the Unites States. That being the case, you should find out where the Democratic and Republican contestants in your state stand on the issue of missile defense before you vote in November. Generally speaking, Democrats have been opposed to it.

In 2004, Barbara Boxer put forward an amendment to prevent the deployment of the first stage of our national missile defense system in Alaska. The amendment failed, and several interceptors, imperfect though they may be, have now been deployed. If North Korea launches a missile at the U.S. in the near future, I'll be glad to have that system in place. Democrats, by contrast, wanted the system to be 100% reliable before it was deployed (I doubt they wanted it deployed at all, but let's take them at their word). As Joseph Biden said:

At this time, we cannot be sure that the actual system would work against a real North Korean missile threat.

No, we can't be sure. But we can be 100% sure that a nonexistent missile defense system will be unable to prevent a nuclear attack from North Korea. The system that is in place has at least some chance of succeeding, and that's precisely why it is there. If North Korea ever does decide to launch at missile at our west coast, I feel sure that even Joseph Biden will be in favor of trying to shoot it down with the interceptors he so adamantly opposed.

With regard to the Boxer amendment, here is how Senators voted in the states where key election contests are being held:

Pennsylvania:

In Pennsylvania, Rick Santorum voted against the Boxer amendment (i.e., he voted in favor of deployment). Pennsylvanians should ask his opponent, Bob Casey, where he stands on this issue. Casey is ahead in the polls.

New Jersey:

In New Jersey, the outgoing Democratic Senator voted in favor of the Boxer amendment (i.e., he voted against deployment). Do voters in New Jersey know where Thomas Kean and Robert Menendez stand?

Montana:

In Montana, Republican Conrad Burns voted against the amendment and in favor of deployment. Where does Democrat Jon Tester stand? Tester leads in the polls.

Missouri:

In Missouri, Republican Jim Talent voted for deployment (not for the Boxer amendment). Where does Democrat Claire McCaskill stand? McCaskill leads in the polls.

Tennessee:

In Tennessee, the outgoing Republican (Frist) voted for deployment. Where do the competing candidates stand?

Ohio:

In Ohio, Republican Mike DeWine voted for deployment. Where does Democrat Sherrod Brown stand? Brown currently leads.

Rhode Island:

In Rhode Island, Republican Lincoln Chafee voted for deployment. Where does Democrat Sheldon Whitehouse stand? Whitehouse currently leads.

For the most part, Republicans want to aggressively pursue missile defense, whereas Democrats want to stand in the way. North Korea just tested a nuclear bomb, and they are working hard on their ballistic missile technology. It won't be long before they have a missile that can reach our shores.

Cast you vote wisely in November.

The Median Family Income Picture is Bright, Too

The Mark Folely affair is a micro-scandal that will be of no lasting importance whatsoever. The state of the American economy, by contrast, is enormously important (both to us and to the rest of the world). In almost every respect, the economy is in fantastic shape, yet the mainstream media still has not taken notice of it. And when they do look at the economy, they search hard for a dark lining on what is clearly a bright and shining silver cloud.

In their presumably well-intentioned zeal to ensure that the poor and the average fellow participate in the economic good times, mainstream media reporters often get it wrong. This is especially true when the issue concerns family income. A story in the Wall Street Journal (from back in March of this year) is typical. The article relates a conversation with then Treasury Secretary John Snow:

Economic output has increased at an annualized pace of almost 4% since mid-2003, and the unemployment rate has fallen to 4.8% from 6.3%. Despite that, polls show more Americans think the economy is worsening than think it is improving.
...
Mr. Snow distributed a fact sheet that showed after-tax income per person, adjusted for inflation, rose 8.2% from January 2001, when George W. Bush took office as president, through January 2006.

Mr. Snow's case relies on averages, which can be skewed by big gains among the wealthiest. Other data suggest the typical family has seen little advance in income or net worth since Mr. Bush took office. Census Bureau data show median family income -- half of families have income greater than the median, half have less -- fell 3.6% from 2000 through 2004. Incomes for the poorest families fell even further. The only group to gain was the family at the 95th percentile -- that is, richer than 95% of all families. Data for 2005 are unavailable.

The story included this helpful little graphic:


This is simply not serious, but I have seen variants of it over and over again. Oh sure (the reporter implies), the wealthy are doing well, but what about the average guy? He's getting screwed, with median income down 3.6%. Some recovery.

The median family income statistic is a pre-tax figure, and President Bush pushed through the Economic Growth and Tax Relief Reconciliation Act of 2001 to enhance after-tax income (which is the only income that matters to your average Joe). John Snow tried to provide some after-tax information to this reporter, but he provided a per-capita (i.e., an average) statistic, not a median statistic. An average value can be pulled up even if the rich alone are seeing their incomes increase. That's what the reporter thought he detected by looking at the median income statistic (which shows a decline). In truth, the story told by the average and the median statistics rarely diverge, but you'll never convince a liberal reporter of that, so it's best to stick with the median just to avoid the issue.

A competent reporter might have asked Snow about median after-tax income if he didn't like the average after-tax income figure. A reporter with an agenda, on the other hand, would just write an article designed to make you think that the middle guy is worse off now. And that's what this reporter did (while ironically noting that most Americans don't think well of their economy).

As a brief aside, you might be surprised that an article like this -- one that pushes a standard liberal talking point -- would appear in the conservative Wall Street Journal, but you shouldn't be. This newspaper is unique in that its editorial page is conservative (which is what you've heard), but their news operation is about as liberal as it gets.

Anyway, back to the issue at hand. First, here is the figure from the Census Bureau that the reporter was presumably citing (found on page 4 of this report):


These figures have been adjusted for inflation and are expressed in 2005 dollars. The first thing to note is that, with the added perspective provided by this chart (compared to the one the reporter prepared for us), it is clear that the median American is doing very well in terms of income. Even after adjusting for inflation (which takes into account higher healthcare costs, higher housing costs, higher eductaion costs, etc.), incomes are as high as they have ever been save for the brief but unsustainable bubble economy that occurred during the latter years of the Clinton administration (a bubble that was based primarily on "irrational exuberance").

You can see from the chart that we had a recession when the stock market bubble burst in 2000/2001 (the attacks of 9/11 didn't help either). And sure enough, median incomes declined between 2000 and 2004 (and then finally started to recover in 2005, as news reports have finally taken notice of). The reporter for the Wall Street Journal report did not have the 2005 information at the time the article was written, and he appears to have correctly interpreted the median income story from 2000 to 2004 (i.e., median incomes were headed down).

Except that, as I said, this is pre-tax median income. What does after-tax median income look like? After all, the tax cut was designed to help offset the effects of a looming recession, and one would like to know what the effect of the tax cut was for the typical person. Democrats often say that the tax cuts only help the wealthy, so it is surely not the case that after-tax median family income started increasing before 2005. No possible way.

Right?

Wrong. Here's a figure from a study by the Joint Economic Committee:


Folks, this is median income for families with two children. This is the "typical guy." And this graph shows the difference between after-tax income with vs. without the tax cut. The typical guy was doing pretty well (after taxes, adjusted for inflation) by 2003. Certainly nothing to complain about. As with all other economic indicators, after-tax incomes were improving way back in 2003 despite the income story that is relentlessly pushed by the media.

I've checked these values against household income figures supplied by the Congressional Budget Office, and a very similar trend is evident there. I found the data through 2003 here and I computed after-tax income for the 40%-60% income group (not technically the median, but close enough):


The exact income figures in this graph differ from the family income figures in the graph above because this one applies to all households (not just families with two children), but the pattern is the same. Household income began to recover in 2003, not 2005. Note that, in this graph, I've also highlighted the bubble economy years (shown in red) just to emphasize that it is silly to be comparing current income values against those artificially inflated values. But that's what reporters do (without noting the bubble), so I guess I will, too.

To reiterate: household income began to recover way back in 2003. Here's a good example of some reporters missing the boat:

Middle-Class Workers Ailing in Census Checkup
Household incomes rose from 2004 to 2005, but earnings fell among full- timers, the agency says. The ranks of uninsured grew by 1.3 million.

By Joel Havemann and Ricardo Alonso-Zaldivar, Times Staff Writers
August 30, 2006

WASHINGTON — The Census Bureau's annual snapshot of economic health in America offered a yellow warning light for the middle class, as an unchanged poverty level and a widening erosion of health insurance coverage tarnished news that household income was finally beginning to rise.

Household income rose from 2004 to 2005 for the first time since 1999, the agency said in its report, released Tuesday.

Well, yes, if you are talking about pre-tax income. A more curious team of reporters might have inquired into after-tax income, which is what really matters. What they would have learned is that after-tax median family income began to increase much earlier -- from 2002 to 2003 (despite the fact that pre-tax income fell over those years). To put this another way, the income recovery was underway for 2 years before any reporter noticed it. In those two years, reporters have consistently told a bleak income story, which has no doubt contributed to the fact that Americans hate their fabulous economy.

I don't have more recent after-tax income figures (I've looked everywhere), but I really wish I did because I am reasonably sure that the income story would look even better now given that various tax cuts were accelerated in 2003. When the 2004 and 2005 figures become available, I'll share them with you because I doubt that any reporter will.

Oh, and about the poverty rate figures that serve as a "yellow warning light" according to these reporters? Here is some perspective from page 13 of that Census Bureau report:


The poverty rate, which is the lower line, has bounced around in a constant zone for about 35 years now (i.e., it's not a warning light), and it showed some slight improvement due to the bubble economy as well. More on that some other time.

For now, my point is that the median family income story -- after taxes -- is brighter than the median family income story the media likes to tell. And the reason it is brighter is because of Bush's tax cuts, the ones that supposedly benefited only the rich (the graphs in this post show that isn't true) and supposedly caused the federal deficit to spiral out of control (also false).

Let's not have a National Enquirer-type election this November. Instead of casting your vote based on what is, in the grand scheme of things, a trivial micro-scandal, cast it based on something important, like the economy or the war in Iraq or the war on terror. And before you vote, consider which party is asking you to focus your attention on the important matters and which would rather have you focus on a trivial micro-scandal.

Update on the Federal Budget Deficit

The Mark Foley micro-scandal can't change the fact that the American economy is going great guns. The 2006 budget deficit figures were just released, and the deficit is not only getting better, it is now relatively small by historical standards. Most news stories don't mention that fact, but this one did:

Budget Deficit Drops to $250 Billion
By ANDREW TAYLOR , 10.06.2006, 01:53 PM

The federal budget deficit estimate for the fiscal year just completed has dropped to $250 billion, congressional estimators said Friday, as the economy continued to fuel impressive tax revenues.
...
The improving deficit picture - Bush predicted a $423 billion deficit in his February budget - has been driven by better-than-expected tax receipts, especially from corporate profits, CBO said.
...
At 1.9 percent of gross domestic product, the 2006 deficit registers far below those seen in the 1980s and early 1990s. The modern record of 6 percent of GDP came in 1983 and deficits greater than 4 percent in 1991 and 1992 drove Congress to embark on a 1993 deficit-cutting drive.

I was amazed to see some historical perspective like this. The standard technique is to cite the absolute size of the deficit to make it sound much bigger than it really is. But what you need to know is the size of the deficit relative to the GDP, and that number is down to 1.9%.

For even more perspective, here are the deficit figures going back to 1976 (negative numbers represent a deficit, positive numbers represent a surplus):

We are now well below the the average budget deficit over the last 30 years, which is 2.6%. And look at that trend over the last 3 years. It's better than anyone predicted as recently as 6 months ago. You can see that we had a surplus during the Clinton years, but everyone agrees that it was an illusion due in large part to the stock market bubble that finally burst shortly after Bush took office. The end of that speculative market, the attacks of 9/11, and the subsequent wars in Afghanistan and Iraq all took their toll on the federal budget. Viewed in that light, it is simply amazing that the deficit is as small as it is now.

Oh, and here is another trend you've probably heard about:



This graph shows the Dow Jones Industrial Average from 1997 up to today. The market peaked in the year 2000, but then the bubble popped. As you can see, though, the market has been on its way back for a while now, so much so that it is in record-setting territory once again.

Let's see, GDP growth is strong, productivity is way up, unemployment is low, family incomes are rising, inflation is down, the stock market is up, and the federal deficit is down. Yet polls show that we are about to put more Democrats into both the House and the Senate. That should fix all of these economic problems, real quick.

Is Tony Snow Responsible for the Bush Bounce?

Bush's popularity has been making a comeback lately, and I have not really had a good explanation it. The recovery started before oil prices began coming down and before new terror threats emerged.

A recent story in the Washington Times (linked today on freerepublic.com) offers one possible explanation:

Presidential ratings rise with new message man
By Joseph Curl
THE WASHINGTON TIMES

Former talk show host Tony Snow took over as President Bush's communications point man four months ago, beefing up the press office staff, honing internal operations and deploying a quick-response strategy.

Now, polls show, the president's approval rating has jumped to its highest level since January.

Could Mr. Snow be responsible for the surge?
...
Even though he says he was "absolutely scared stiff because I didn't know what to expect, I'd never done anything like it," Mr. Snow says he has fallen in love with the job he has performed since May.

In earlier posts, I discovered something that this reporter could not have known: Bush's popularity bottomed out in May and has been climbing ever since (the approval rating is the lower blue line):

These numbers are based on the average of many polls, making the sample size close to 20,000 in some cases, so the results are virtually free of statistical noise (such that even small changes are meaningful). Using this method, we can pinpoint the low point: it occurred in May.

Tony Snow started his job as the White House Press Secretary in May, and Bush's approval ratings have been rising ever since. There may be no connection, but the timing supports Joseph Curl's theory.

If the theory is right, it is presumably not the case that Americans are watching Tony Snow and thinking better of Bush. Instead, Tony Snow may be influencing how reporters cover the stories of the day. Although we can't know for sure, it seems reasonable to suppose that Bush's popularity has gone up as Tony Snow has induced liberal reporters to present the news in a less biased way.

Update on the New Jersey Senate Race

The New Jersey Senate race is interesting because, as I explained here, it is the one seat held by a Democrat that a Republican seems likely to take away.

At least it seemed that way until today. All of a sudden, the RealClearPolitics average poll has the Democrat Menendez up over Kean by a sizeable 3.4%. And, as I explained here, the average poll is the only one that matters (not individual polls). A 3.4% lead in an average poll is nothing to sneeze at.

What's puzzling about this apparent turn of events is that Kean was already running ahead in the polls when this news story broke a few days ago:

N.J. Senator Accused Of Kickbacks
Taped Phone Calls Could Damage Menendez's Campaign

TRENTON, N.J., Sept. 29, 2006

(AP) A childhood friend and close political adviser of Democratic Sen. Robert Menendez allegedly used the senator's name to pressure a psychiatrist to hire a doctor for local government work, according to a transcript of the telephone call, part of which is mentioned in federal court documents.

My first thought was that news like this would further cement Kean's lead. They thought so over at the NRO Corner as well, with an entry entitled "Senator Kean, Welcome to Washington." I shared this thought by John Podhoretz:

"U.S. Sen. Robert Menendez's closest political adviser was secretly recorded seven years ago boasting of political power and urging a Hudson county contractor to hire somone as a favor to Menendez, according to a transcript obtained by The Star-Ledger." Ah, New Jersey. The close political adviser is Menendez's best friend since childhood -- but Menendez didn't know he was going around muscling people for crony jobs, according to his campaign. Bye bye, Menendez.

Much to my surprise, new polls suddenly show Menendez in the lead, and several of the recent polls were taken after this supposedly damaging news broke. This puzzling turn of events suggests that my recent dismissals of Democratic chances of taking the Senate should have been made with slightly less confidence.

Averaged polls don't usually show such volatility, so I took a closer look at the data. RealClearPolitics usually averages over the last few polls only, and those polls are often based on only 600 or so respondents each (which means that the total sample size in the averaged poll often falls between 2000 and 3000). That sounds like a lot of people, but to get stable numbers, you need an even bigger sample size, so I took a closer look at this race over the course of 2006 using substantially larger sample sizes. Here is what I discovered:


The numbers for the month of September were averaged over all of the polls from that month. Earlier polls were averaged over multiple months (there were fewer polls back then, so I had to average over a larger block of time to get a large sample size). What you see is that this race has been a statistical dead heat all year long, with a very slight advantage for Menendez evident for a number of months now. The difference between the two candidates lies with the margin of error even for these very large sample sizes (generally between 6000 and 7000 people), but the consistency of the slight lead by Menendez makes it hard to argue that he is running behind. It's clearly a statistical dead heat, which differs from what I previously thought.

Consistent with this new analysis, TradeSports.com now shows a virtual dead heat in the betting (52.0 for Menendez vs. 51.5 for Kean as of this moment).

What all this means is that even averaged polls from RealClearPolitics, while being vastly better than individual polls (which tell you next to nothing), are subject to a bit too much variability as well. I'll continue to bring you updates based on their averaged polls, but I may compute my own averages (as I did here) from time to time in an effort to get a clearer picture.

For the moment, the point is that New Jersey is a toss up. If anything, Menendez has a very slight lead.

UPDATE: the traders at Tradesports.com seem to have caught on. A few days ago, they were betting on Kean. This morning it was about even. As of 1:30 P.M. on the west coast, it's 53/46 in favor of Menendez.

Walter Pincus is Torturing Me

Walter Pincus has a new column in the Washington Post entitled "Waterboarding Historically Controversial." In it, he says:

Twenty-one years earlier, in 1947, the United States charged a Japanese officer, Yukio Asano, with war crimes for carrying out another form of waterboarding on a U.S. civilian. The subject was strapped on a stretcher that was tilted so that his feet were in the air and head near the floor, and small amounts of water were poured over his face, leaving him gasping for air until he agreed to talk.

"Asano was sentenced to 15 years of hard labor," Sen. Edward M. Kennedy (D-Mass.) told his colleagues last Thursday during the debate on military commissions legislation. "We punished people with 15 years of hard labor when waterboarding was used against Americans in World War II," he said.

How shocking. A technique that was punishable by 15 years of hard labor not so very long ago is now perfectly acceptable according to our apparently demented commander in chief! Pretty much ends the torture debate, don't you think?

Right.

First, the administration proposes to use waterboarding only on high-level combatants who may have information that could save thousands of innocent lives. Second, and more to the point, Yukio Asano was not sentenced to 15 years for using that one technique, as Ted Kennedy implies. At least not according to the information I found about 15 seconds after entering the words Yukio Asano 15 years into a google search:

Defendant: Asano, Yukio

Docket Date: 53/ May 1 - 28, 1947, Yokohama, Japan

Charge: Violation of the Laws and Customs of War: 1. Did willfully and unlawfully mistreat and torture PWs. 2. Did unlawfully take and convert to his own use Red Cross packages and supplies intended for PWs.

Specifications: beating using hands, fists, club; kicking; water torture; burning using cigarettes; strapping on a stretcher head downward

Verdict: 15 years CHL

Why must liberal critics constantly resort to the tactic of lumping together techniques that cause real physical injury, like beating someone with a club or burning them with a cigarette (which everyone agrees is torture), with harsh techniques that do not cause permanent physical injury (like waterboarding)?

It's fine to draw the torture line short of waterboarding if you are so inclined, but it's not reasonable to resort to silly little gimmicks to avoid having a real debate about it.

Pincus goes on to say:

In the post-Vietnam period, the Navy SEALs and some Army Special Forces used a form of waterboarding with trainees to prepare them to resist interrogation if captured. The waterboarding proved so successful in breaking their will, says one former Navy captain familiar with the practice, "they stopped using it because it hurt morale."

Were our Navy SEALs being tortured by the Navy during these sessions? No. They stopped using the technique because it worked so well that it hurt morale. You haven't tortured someone merely because you have wounded their pride.

The fact that the Navy used the technique on our own soldiers goes a very long way toward making the case that waterboarding is not torture. In fact, it suggests to me a new test for when we should apply that nebulous word. Specifically:

It's not torture if we are willing to use it on our own soldiers while training them to fight a war.

That simple test would prevent the actual torture of any al Qaeda detainees, but it would still allow for the use of harsh interrogation techniques that could save thousands of innocent lives.

What's the harshest interrogation technique that liberal critics would regard as being acceptable? They will never say, and that speaks volumes. They are in this argument to play a political game, not to seriously consider the question of how to interrogate high-level al Qaeda detainees.

Update on U.S. Senate Races: Averaged Polls and Tradesports.com

In a recent post, I explained what you need to know about polls. The upshot is that individual polls are so variable that you cannot really tell very much from them. A much better approach is to rely on the average of multiple polls. If you have ever been confused by apparently contradictory polls, read my previous post to see what is probably going on!

Individual polls are good for making apparent news out of random error, which is what the mainstream media does all the time. Averaged polls, by contrast, are good for obtaining a close approximation to the truth. RealClearPolitics.com is the place to go to get averaged polls.

Another useful approach to forecasting the future is to rely on the monetary bets people make. There is wisdom in those bets, presumably because when people put money down, they attend to all of the relevant information (polls, the amount of advertising money the candidates have, etc.). Tradesports.com is the place to go to get information on how people are betting in the various Senate races.

In this post, I'm going to give you an update on the Senate races that are of most interest because the seat may change from one party to the other. I'll do that by giving you the latest information based on the average poll from RealClearPolitics as well as the latest information based on the bets people are making from Tradesports.com.

The following table gives you everything you need to know. The states of interest are listed in red or blue. Red means that the Senate seat is currently held by a Republican, whereas blue means that it is currently held by a Democrat. Look down at the table and you can immediately see that most of the action is in states that are currently red. I then show who is ahead in the Senate Race according to the average of multiple polls from RealClearPolitics and according to the betting at Tradesports.com:



So, the fist state (Missouri) is in red because it is currently held by a Republican. The Senate race in that state is very close, and it is the only one where the two methods of forecasting the future disagree. The averaged poll says that the Democrat (McCaskill, in blue) has a slight lead, whereas the betting slightly favors the Republican (Talent, in red). That's why I have question marks off to the left.

Montana, Ohio, Pennsylvania, Rhode Island and Tennessee are all red states that both methods predict will become blue states (so I put a blue asterisk off to the left). New Jersey is a blue state that both methods predict will become a red state (so it gets a red asterisk).

Thus, 5 weeks out from the election, it looks like the Democrats will pickup 5 Republican seats (those 5 blue asterisks), and the Republicans will pick up one Democratic seat (the one blue asterisk). That's a net gain of 4 for the Democrats, with one race in doubt (signified by the red and blue question marks). The Democrats need 6 seats to regain control of the Senate, and it still looks to me like there is little chance of that happening. Instead, they'll get 4 or 5.

I also included the results of 2 other Senate races that have a lot of interest right now, Connecticut and Virginia. Connecticut looks like a done deal: Lieberman is going back to the Senate. Virginia looks like a done deal as well, with Allen holding on to his seat.

The Mark Foley scandal shows that anything can happen between now and election day, but the best predictors we have right now say that the Republicans will hold on to the Senate. Rest assured that I will keep you posted as the election draws closer.

Ann Althouse asks about Same-Sex Marriage

Ann Althouse asks if there is a rational basis for opposing same-sex marriage (via Instapundit). I think there is.

I'm not saying that I'm opposed to same-sex marriage, but I am saying that those who are opposed to it cannot be described as being irrational.

I am sure that others have said this better than I can, but one rational argument against same sex marriage derives from the rational basis for marriage itself. Marriage came into being (I assume) because people eventually realized that the common sexual attraction between a man and a woman often yields children. Creating a socially enforced bond between two people who produce a child is advantageous to society in the sense that it is much better to have a man working to improve the lot of his family unit than it is for him to be doing what comes more naturally (which is to sleep with every woman he finds to be physically attractive).

So, according to this argument, the rational basis for marriage is not sexual attraction per se. It is, instead, the fact that sexual attraction between men and women often results in children, and women with young children are in a vulnerable economic situation. Moreover, the rational basis for marriage involving 2 people (not 3 or 4) is that 2 people -- and only 2 -- produce a child.

To say that opposition to same-sex marriage is irrational is to say that the basis for marriage is sexual attraction per se (i.e., marriage exists to strengthen the bond two people who are sexually attracted to each other). But, according to the argument presented above, sexual attraction, per se, is not the basis for marriage.

If children figure into the calculus of marriage, doesn't that mean that we should ban marriage for older heterosexual couples who are beyond their child-bearing years? It would not be irrational to do so, but there would be no point to it either. Any cost associated with allowing heterosexual couples to marry even if they can't have children is outweighed by the cost associated with having to rigorously test every couple planning to get married just to prove that they really are capable of bearing children.

No such rigorous testing is needed to prove that same-sex couples can't produce children. That being the case, any cost associated with allowing them to get married (if, indeed, there are any such costs), is not outweighed by the cost of testing to see if they can produce children. No special expertise is required to appreciate the fact that these two people are not going to produce children.

I am not saying that gays should not be allowed to marry. I am saying that opposition to gay marriage is not irrational. And I'm admitting that denying two 80-year-olds the right to marry would not be irrational, either. But it would be pointless because there are no costs for allowing it.

It might also be pointless (but not irrational) to deny gays the right to marry if there are no costs associated with allowing them to marry. Whether or not such costs exist is the crux of the issue.

Muqtada al Sadr: Our Unintentional Ally?

In light of increased sectarian violence in Iraq, many have wondered why we don't find a way to disarm the militias that are wreaking havoc in Baghdad. It's a reasonable question, but I think there might be a pretty good explanation for why we don't.

First, it's worth considering who the militias are and who they aren't. This description taken from a recent Defense Department report opened my eyes a bit:

Unlike the Kurdish and Shi'a militia groups, Sunni Arabs do not have formally organized militias, but rely on neighborhood watches, Rejectionists, and, increasingly, al-Qaeda in Iraq (p. 30).

So, when you say "disarm the militias," you really mean "disarm the Shiites only." The Sunnis do not have organized militias, and the Kurd militias are not fighting anyone (because the Kurdish areas are relatively peaceful).

Here is more from the report (note that JAM stands for Muqtada al Sadr's Shiite militia):

Sunni and Shi'a extremists, particularly rogue JAM elements and al-Qaeda in Iraq, are, as noted, interlocked in retaliatory violence and are contesting control of ethnically mixed areas to expand their areas of influence. Throughout the past quarter, rogue JAM members continued a campaign of overt executions and mass kidnappings of Sunni civilians. At the same time, Sunni extremists continued to respond by carrying out large scale and mass-casualty bombings of Shi'a gatherings and culturally significant sites (p. 34).

So, why don't we eliminate Muqtada al Sadr's militia? Because that militia has become a serious threat to the stubborn Sunni-based insurgency that we've been fighting for the last 3 years. As such, taking out the militias would strengthen the hand of our main enemy. Thus, in an odd way, Muqtada al Sadr may now be our unintentional ally.

US forces have not been able to fully quell the insurgency, and the Sunnis appear to believe that they have more staying power than we do. One advantage they have in this fight is that they mingle with civilians to launch their strikes, whereas American soldiers try to fire only on armed combatants while sparing the lives of near-by civilians. It's an almost impossible task.

Muqtada al Sadr is not similarly constrained by ethical considerations. His militia summarily executes large numbers of young Sunni males, seemingly every day. I believe that this is an example of his militia in action from just a few days ago:

BAGHDAD - A total of 50 bodies were recovered by Baghdad police in various parts of the city over the 24 hours to Sunday evening, an Interior Ministry official said. Many had been tortured and most were bound and shot in the head.

That kind of brutality has an effect on Sunnis that American soldiers could never have. In fact, the Sunni community is beginning to live in fear as a result of these on-going attacks. Moreover, all of a sudden, the Sunnis are not quite as interested in seeing US forces leave the country, which is a rather amazing development. A recent report that accompanied the latest poll from Iraq conducted by WorldPublicOpinion.org said this about the withdrawal of American troops from Iraq:

Shias show a growing a sense of urgency, with the numbers calling for withdrawal [of American troops] in six months rising from 22 percent to 36 percent. On the other hand, the Sunnis’ earlier overwhelming eagerness for withdrawal has moderated, with the percentage calling for withdrawal within six months dropping from 83 percent to 57 percent. Among those living in Baghdad support is even lower at 24 percent.

A majority of Sunnis still wants us out of the country, but the trend is unmistakable. And that apparent change in Sunni attitude may be one reason why the Iraqi government is not yet ready to deal forcefully with Muqtada al Sadr. Only he seems to have the power to influence Sunni thinking, and to disarm him would be to disarm a major threat to the Sunni opposition we face.

Consider this recent article about a new plan in Iraq for ending sectarian violence:

Iraqi lawmakers endorse peace plan
...
Sunnis accuse the Shiite-led security forces of turning a blind eye to killing of Sunnis by Shiite militias, some of which are linked to parties in the government. Sunnis have accused al-Maliki, a Shiite, of being hesitant to crack down on the militias.
...
Shiites, meanwhile, accused Sunni parties of links to terrorists after a bodyguard of a Sunni party leader, Adnan al-Dulaimi, was arrested Friday by U.S. forces and accused of plotting al-Qaida bombings.

The Sunnis probably have it about right (i.e., the Shiite-led security forces probably are "turning a blind eye to killing of Sunnis by Shiite militias"), but their problem is that the Shiites have it about right, too (i.e., the Sunnis really are in league with terrorists in their efforts to regain control of the country).

Before the government cracks down on Muqtada al Sadr, I suspect that the Sunnis running the insurgency are going to have to finally make real concessions. The are not doing that yet, but their attitude may be changing as their situation deteriorates. Before too much longer, they'll either make some concessions or they'll be begging US troops to stay.

Finally, on a related note, I remembered a StrategyPage report I read on Instapundit from way back in January of 2005. StrategyPage said this:

Can the anti-government forces in Iraq win? Some pundits think so. But do you really think the Shia and Kurds will allow Saddam's thugs to bully their way back into power? The Kurds and Shia Arabs have 80 percent of the population, control of the oil, and American troops to back up their efforts. Iraqis indicate, to anyone who will listen, that they have no intention of folding under Baath pressure, and a growing desire to come down hard on the Sunnis who support the violence. The Kurds and Shia have names, because Saddam's thugs didn't wear masks when they ran things for three decades. Guess who is going to lose? But that thought is what is driving the resistance. The Baath Party thugs know what they will have to face eventually, if they don't regain control of Iraq.

Well, the Baathists are not going to regain control of Iraq. And so long as the Iraqi government (backed by US forces) does not eliminate the threat posed by Muqtada al Sadr, they are going to die, too. Their only hope may be to make a deal with the government.

I'm not endorsing the idea of allowing al Sadr's militia to execute young Sunni males on a large scale. I am just describing what I think I see happening over there. In a war, you are sometimes faced with seemingly impossible choices. The choice here may boil down to this:

1. disarm Muqtada al Sadr's militia to protect innocent Sunnis, which will also serve to strengthen the Sunni-led insurgency (which, in turn, may cause us to lose Iraq)

2. stand aside while Muqtada al Sadr's miltia summarily executes large numbers of Sunni civilians until the Sunni-led insurgency is finally ready to stop the fight.

That's my theory about why we don't disarm the militias that are wreaking havoc in Baghdad. I'll update that theory if new developments in Iraq prove it to be wrong. For the time being, though, I believe that the Iraqi government (and the US military) may have chosen option #2 because the Sunnis have chosen to fight rather than cooperate. Thus, in very real way, it's the Sunnis themselves who have made the choice. They need to stop the fight to stop the slaughter of their own people.

Bush's September Job Approval Rating

In the post below this one, I explain why you need to average over many polls to get a clear understanding of what is going on. That rule applies to Bush's approval rating, too.

I averaged the results of many polls conducted each month since December of 2005 so that trends in the approval rating could be observed free of noise due to random error (noise that can really distort the picture). That is, the rating for each month is an average of many polls found here, not the results of any one poll. I presented this graph a couple of weeks ago, and I'll now present the updated graph (which, unlike before, has complete data for the just-completed month of September). The approval rating is shown in blue:

As I said, each point on the graph is an average of many polls, so the trends are real. Over the last few months, we can say with high confidence that Bush's approval rating has been steadily (albeit slowly) climbing. I'll let you know what happened in the critical month of October when the time comes.

Be sure to scroll down to the post below to see why you should always let your mind be guided by averaged polls (not individual polls).

What You Should Know about Polls

There is a new story on the upcoming Senate races that starts out like this:

Polls: Democrats in striking distance for Senate
Several incumbent Republicans trailing with five weeks until election

NEW YORK - Five weeks out from the midterm elections, MSNBC/McClatchy polls, conducted by Mason-Dixon in eight states, show Democrats are in striking distance of taking control of the Senate. The Democrats are very likely to gain several Senate seats with some races still rated as toss-ups.

What most people don't realize is that the polls presented in this story don't tell you very much at all. In this post, I'm going to explain why that is.

First, it is important to distinguish between two kinds of polls that are often in the news. The first is what you might call a number poll, such as a poll that measures Bush's approval rating. In a poll like that, we are mainly interested in a single number, such as the fact that the President's approval rating is at 41%. The second is what you might call a difference poll (such as a Senate race between a Democrat and a Republican). Here, the question of interest is: "who is ahead and by how much?" In a difference poll, you don't really care about each candidate's exact level of support. For example, you wouldn't care that Candidate 1 has 45% support because that information alone doesn't tell you what you need to know. What you need to know is whether that value is greater than or less than Candidate 2's level of support. If Candidate 2 has 40% support, then supporters of Candidate 1 will be happy to know that their candidate is enjoying a 5% lead. That 5% value is a difference score (i.e., 45% - 40%), and that difference score is what we are really interested in knowing.

Why does it matter if we are talking about a number poll (as in a poll about the President's approval rating) or about a difference poll (as in a poll about a Senate race)? Well, I am sure that you've heard a reporter say something like "the poll had a margin of error of 3.5%." This is where it matters. Stick with me because it's easy, and it's worth knowing.

In the news story I linked to above, it says that the polls had a sample size of 625 people, which resulted in a margin of error of 4%. I'm now going to show you the results of 20 simulated polls, each one based on a sample size of 625 people. These are simulations from a computer program that used a random number generator, but the results are just like what you get when you use real random sampling to conduct a poll. Each of my simulated polls had a sample size of 625, and in every one of them, I set Candidate 1 (say, George Bush) to have 49% support and Candidate 2 (John Kerry) to have 44% support (with the rest being undecided). That is, the underlying reality remained constant in all 20 polls, so the results of the polls differed only because of random error (just as would happen in real polls). Random error can never be avoided, except by massively increasing the sample size. But the sample size is fixed at 625 in this case, and here are the results (I changed this to an image from an earlier table that had formatting problems):


What these simulated polls show is how much you would expect real polls involving 625 people to differ just due to random chance. The first thing to note is that the average over all 20 polls comes very close to capturing the underlying reality. That underlying reality is 49% vs. 44% (a 5% lead for Bush), whereas the average suggests 49.6% vs. 43.5% (a 6.1% lead for Bush). In other words, the average gets it about right. It's always that way.

The second thing to note is that the individual polls are simply all over the place. Some of the polls actually show Kerry in the lead. Others show Bush up by 12 when, in fact, he is up by only 5. This occurs even though the underlying reality did not change at all from poll to poll. Does this kind of confusing variability ring a bell? It should because this is how it was in the run up to the last presidential election. The underlying reality did not change much as the election approached, but the polls were really bouncing around. Well, they seemed to be, but it was mostly random error.

Bottom line: The truth is found in the average poll, not in the individual polls.

Because the margin of error is 4%, you'd expect the individual values for Bush to fall in the range of 49% (the true value) plus-or-minus 4% (the margin of error). That is, the values should range from 45% to 53%, and that's the actual range that you see across the 20 simulated polls. So, that's what the margin of error tells you. That is, it tells you the likely range into which the poll numbers will fall due to random error alone. And if this were an approval rating for Bush rather than a contest between two people, that's all you'd really need to know.

In reality, Bush's approval rating is about 41% right now. That means that poll results should range from about 37% to 45% (i.e., 41% plus-or-minus about 4%). Sure enough, if you look at the polls conducted during the month of September, Bush's approval rating ranges from a low of 37% to a high of 44%. It's all random error. But that won't stop reporters from writing stories like this:

Sex scandal, Iraq book take toll on Bush, GOP
NBC/WSJ poll: Bush back in 30s; Dems favored for control of Congress

In the poll, Bush's job approval rating is at 39 percent among registered voters, a drop of three points since September, when his rating had increased to its highest level in months after he gave a series of speeches on national security leading into the fifth anniversary of the Sept. 11 terrorist attacks.

This is an example of what I call "making news out of random error." There is no news here. The approval numbers are just bouncing around as they always do (and as they must, unless the same size is huge). Bush's approval might be dropping, but there is really no way to tell from this poll because you'd often get results like this even if his approval rating were still at 41%.

Anyway, back to the other kind of poll, which involves a contest between two people (e.g., Bush vs. Kerry). In that case, no one really cares about the individual numbers. What we care about is who is ahead and by how much. That is, we care about the difference between the two values (which I show in the column labeled "diff"). Here is the key point I want to make: the expected variability in the difference score is much greater than the margin of error. In fact, generally speaking, it is about double the advertised margin of error. That being the case, in my simulated polls, we'd expect the difference scores to fall in the range of 5% (the true difference) plus-or-minus a whopping 8% (i.e., from -3% to 13%). And that is about what you see in the simulated polls above (although the biggest difference was 12%, not 13%, for these 20 polls).

The point is that in a race between 2 people, the difference score is what you want to know ("who is ahead and by how much?"), and the margin of error for that difference score is about double the margin of error that is advertised. That's how sloppy an individual poll based on 625 people is. The difference between two candidates must be outside 2 times the margin of error (not outside the margin of error alone) before the results are convincing. And it's not a whole lot better when you sample 1000 people (in which case the margin of error drops to about 3.2%, which means that the margin of error for the difference is 6.4%). That's why, if you want to know the truth, you have to rely on the average of many polls no matter how intoxicating individual polls might be (and they are undeniably intoxicating). Individual polls suffer from unavoidable random error, no matter how many "adjustments" pollsters make to try to come closer to the underlying reality. A much better approach is to average over polls. Look again at the table above. You can see that, unlike the individual polls, the average of 20 polls gives you something close to the truth. That's the take home message.

So, if you want to know where the individual Senate races stand right now, ignore that news story I linked to above and get the information about averaged polls from RealClearPolitics. I did that for you here, and I'll be updating the information soon. Another reasonable way to get a close approximation to the truth is to look at Tradesports.com where people actually place bets on the outcome. That's also a lot better than looking at any one individual poll. I'll soon compare the results from averaged polls to the betting at Tradesports.com. I expect to find complete agreement, but we'll see.

Should We Set a Timetable for Withdrawal or Send More Troops?

In a new column, Jack Kelly argues that the time has come to set a timetable for withdrawal from Iraq. His reasons are twofold:

1. "First, it's plain we've worn out our welcome" (i.e., polls show that the Iraqis don't like us and want us to leave)

2. "Second, there is little likelihood al Qaida could take over when we leave" (i.e., polls also show that the Iraqis hate al Qaeda).

I like Jack Kelly, but he really missed the boat here. First, the idea that we have "worn out our welcome" suggests that we were once welcomed by the Iraqis, but now things have changed and they want us to leave. As I explained here, the Iraqis (excluding the Kurds) wanted us to leave pretty much from the very beginning. As such, it is clear that Kelly's views are being overly influenced by the results of a recent poll, which he mistakenly reads as revealing some sort of new attitude on the part of the Iraqis. That suggests to me that his interest in Iraq is a bit too superficial for him to be taking such a strong stand on a critical issue like setting a timetable for withdrawal.

Second, Kelly shows not the slightest awareness of the main threat facing Iraq right now, which is the sectarian violence that has worsened considerably since al Qaeda bombed the Golden Mosque in February. Al Qaeda may be hated, and they may be on the ropes, but they are still very effective in sewing the seeds of chaos. They proved that. And if we leave too soon, they are fully capable of ensuring that Iraq's experiment in democracy fails (they don't need to "take over" Iraq to do that).

Kelly also cites poll results suggesting that the Iraqis believe that American troops are the cause of much of the violence, and he seems to agree with their assessment. If that were actually true, I could understand why some might think that it is a good idea to start packing our bags. But it's a ridiculous idea, one that can be seriously entertained only if you ignore the fact that sectarian violence is the most serious threat facing Iraq (as Kelly conveniently did). The idea that Muqtada al Sadr executed 60 Sunnis over the weekend because he saw some American troops driving around in a Humvee and then flew into a fit of anti-Sunni rage makes no sense. He ordered his militia to kill those Sunnis because he hates them for what they did back when Saddam was running the show in Iraq.

Make no mistake about it: like Jack Kelly, al Qaeda wants us to set a timetable for withdrawal. In fact, as implied by the recently released NIE report (did Kelly read it?), setting a public timetable for withdrawal is exactly what al Qaeda needs to bring that wretched terrorist organization surging back to life right when they are on the brink of extinction. It's a terrible idea.

If we should not withdraw, should we send more troops instead? Powerline links to a column by First Lt. Peter Hegseth, who served as an infantry platoon leader and civil-military operations officer in Iraq with the 101st Airborne Division. He supports our mission in Iraq but points out that more troops are needed to keep the peace. His column includes these two compelling paragraphs:

In due time, the Iraqi Security Forces will take over Samarra, but they are not ready yet. If the Americans left today, the Iraqis would be co-opted by the insurgents--who are utterly ruthless, willing to kill family members of policemen or decapitate Iraqi soldiers to preserve disorder. It will take time. Both the Iraqi Army and Samarra Police need to get bloodied a bit and bounce back, proving their strength to the people. They will eventually be ready, but until then, security belongs to us.

I also understand calling for more troops is contrary to conventional thinking inside government and the military. Supporters of the current approach argue sending more troops would further inflame anti-American sentiment, incite more violence and retard independent progress. My experience suggests otherwise. American troops are tolerated, even welcomed when they effectively provide security; but their presence is cursed when it does not accompany progress. Violence persists not because American troops are present, but because our presence is futile. Many local leaders asked us, "How come the most powerful country in the world cannot defeat local criminals and thugs?" They suggested our failure was part of a larger conspiracy to keep the Iraqi people suffering.

He makes a stronger case for setting a timetable for expansion than Jack Kelly does for setting a timetable for withdrawal. One's opinion on this matter hinges on the capabilities of Iraqi security forces. First Lt. Hegseth thinks they are just not ready, whereas Jack Kelly seems to share the faith that the Iraqi people have in their own security forces (polls show that they trust their capabilities). My own thinking on this is guided mainly by what the Defense Department says about it, so I'm really happy to have stumbled across their report. The report shows that the sheer number of military and police forces has reached the long-stated goal of 270,000:

But what are their capabilities? The armed forces are doing very well now. The first two bars in the next graph tell that story. In June of '05 (the first bar), not many units were capable of taking the lead, which is the green segment of the bar. In August of '06 (the second bar), many more are able to do that:

That represents real progress, and it seems reasonable to assume that much more progress will be made in the next year. As you have probably heard, though, the police have not made similar progress. Again, the first two bars tell the main story:

There are more police now than there were in June of '05 (i.e., 27 vs. 13), but out of 27 units, only 5 are in the lead. They clearly have a long way to go. Progress is evident (only 2 police units were in the lead a year ago), but it is painfully slow. Part of the problem is that insurgents have been managing to infiltrate the police, which makes the task that much harder (obviously).

Police training is a serious problem, but look at this summary graph of progress in Iraq in the last year:

In October of '05, Iraqi security forces did not have the lead anywhere. Now, they are in the lead in more than half the country (though not in the Anbar province, which is where most of the trouble lies). That is progress, and I believe it reflects a lot more progress than most Americans realize.

As I have said before, we are going to leave Iraq a violent place. About that I have no doubt, even though I sincerely wish it were not true. But, for strategic purposes, the key is to leave when Iraqi security forces can keep that violence from spiraling out of control to the point where their experiment in democracy fails. For that to happen, the police need to receive better training, and you can go to this White House news release to see that more attention is being paid to this problem:

A Major Goal Of 2006 Is To Accelerate Training Of The Iraqi Police. The Iraqi police still lag behind the Army in training and capabilities. One problem is that some police units have been disproportionately Shia - and there have been reports of infiltration of the National Police by militias. We are taking a number of steps to correct this problem:

Just follow the link to read more about the specific steps that are being taken. For now, my point is just this: given the evident progress the Iraqis are making overall, and given the renewed attention being paid to the lagging progress of the Iraqi police force, I do not currently favor increasing troop levels in Iraq. But the argument for doing so is a lot stronger than the argument for setting a timetable for withdrawal.

A Look at September in Iraq

I haven't seen much coverage of the level of violence in Iraq for the just-completed month of September, but that will probably come soon enough (most likely in the form of hysterical claims about how violence is spiraling out of control).

September was a typically violent month, one that was very much like the two preceding months. As usual, I obtained the relevant statistics from Iraq Casualty Count, which tabulates the number of deaths reported by the media. They don't record every civilian death by any means, but their method allows you track trends in a meaningful way. To my surprise, I discovered that there is another authoritative report that provides similar statistics. Mario Loyola at the NRO Corner linked to the Report on Stability and Progress in Iraq, which is released quarterly by the Defense Department. Based on that report, he also challenges Woodward's preposterous claims about Bush keeping "secret" attack statistics. As Loyola notes, the statistics are right there in the Defense Department report, which is available on the web. This just further underscores my emphatic claim (here, here and here!) that the information is most certainly not top secret. Loyola also reads the Defense Department report as saying that attacks against coalition troops per se have actually diminished because the increasing attack statistics that Woodward cites actually refer to attacks of all kinds ("...on Coalition forces, the ISF, the civilian population, and infrastructure"). Attacks on Coalition forces per se may have decreased overall. If so, that also reinforces my claim that, if anything, the insurgency is weakening. I'll show a few graphs from the Defense Department report as I review the September information supplied by Iraq Casualty Count.

First, a technical note: the Iraq Casualty Count statistics for September include 338 deaths that were reported at a morgue in Karbala dating back to January 1. They don't usually do this (i.e., they don't usually include deaths that occurred in other months), and it's too bad they did it this time because if you aren't careful (as a reporter might not be), you could easily get the impression that September was considerably worse than the previous months. Since the vast majority of these 338 morgue deaths did not occur in September, I excluded them from my analysis.

In the first set of graphs, I computed casualties for two different time periods that were defined by the bombing of the Golden Mosque in Samarra back in late February. Period 1 is the 7 months before that (up to and including February), while period 2 is the 7 months since then (i.e., March through September of this year). Here are the average number of casualties per month in 4 relevant categories for those two time periods:

As was already clear from my prior analyses, sectarian violence against civilians worsened considerably after the bombing of the Golden Mosque. Those casualties jumped from an average of about 500 per month before the bombing to almost twice that after the bombing. Al Qaeda bombed that mosque to achieve precisely that result, and it's too bad that it worked. Al Qaeda wanted to incite sectarian violence so that Americans would become demoralized and withdraw prematurely from Iraq. That would have happened had John Kerry won the last election (i.e., like Spain, he'd happily take his marching orders straight from al Qaeda), but it is unlikely to happen while Bush is president.

The bombing of the Golden Mosque marked a turning point in sectarian violence, but it did not mark a turning point in the insurgency. US casualties have remained essentially stable (decreasing, if anything), and the same holds true for Iraqi security forces. Few in the hysterical media appreciate this.

Here are the same figures presented in monthly detail so you can see what happened in September (the red bars represent the months after the bombing of the mosque):



Civilian casualties were high in September (slightly higher than they were in July and August, both of which were bad months). In fact, the last 3 months suggest that violence against civilians has edged up compared to the preceding 4 months, though we'll need wait a bit longer to see if that is a real trend. I hope it isn't because any increase will be astronomically magnified in media reports. They don't hope for bad news, per se, but like most liberal critics, they are excited by any opportunity to savage their own president in a difficult time of war. Unfortunately, such an opportunity is provided whenever casualty statistics go up. I think that Democrats might win more elections if they used such opportunities to weigh in with constructive criticisms, but maligning the president and suggesting that we obey al Qaeda (i.e., suggesting that we set a timetable for withdrawal) doesn't really fall into that category (at least not in my book).

I've argued in prior posts that the mosque bombing ushered in a higher, but essentially stable, level of sectarian violence, not an out-of-control spiral of violence. The civilian casualty data shown in the graph above supports this analysis, and a graph from the Defense Department report that I just found confirms it:



You can see the effect of the bombing of the Golden Mosque in this figure (just as you can in the civilian casualty figure above). I'm glad to see some agreement on this because it increases confidence that the analysis is correct. The overall casualty figures are higher in the Defense Department report, but the trends are the same.

As I have said over and over again, the bombing ushered in a new level of violence, not an out-of-control spiral (which is how the New York Times once described it). That said, it would not be hysterical to suggest that the level of violence against over the last 3 months is higher than the preceding 4 (i.e., July may have marked another step up in the level of violence). While it's certainly not an out-of-control spiral, it would be insane to suggest that things are getting better for civilians.

Another figure from the Defense Department report seems to suggest that Iraqi Security forces are suffering from increased casualties, whereas the numbers from Iraq Casualty Count do not show that trend (their figures go slightly the other way). Here is how Mario Loyola at the NRO Corner describes it:

And in fact, as the following page shows, attacks against Coalition forces have dropped since the summer of 2004, while casualties of Iraqi Security Forces have increased dramatically as ISF have moved to take main responsibility for providing security in the country ( p.32).

And here is the relevant casualty figure from the Defense Department report:


Contrary to what Mario suggests, I don't think that this graph actually shows that Iraqi security forces are taking more casualties as they assume more security responsibilities. That is, as Mario noted, this graph includes Iraqi civilians, and we already know that such casualties have greatly increased of late. In fact, the last two bars showing an increase in the number of casualties mostly corresponds to what I called "Time Period 2" above (i.e., the period of time after the mosque bombing). So, I think the best information still suggests that while civilian casualties have gone way up and remain high, Iraqi security force casualties have come down a bit (along with American military casualties). This is important because if both American casualties and Iraqi security force casualties are edging down, it seems very unlikely that the insurgency is gaining in strength. Instead, the insurgency seems to be gradually weakening. It is critical for people to know whether the insurgency is gaining in strength (as the media hysterically suggests from time to time) or decreasing in strength. Although no one seems to realize it, the relevant evidence suggests a weakening insurgency.

That's the picture from Iraq for the month of September. My main purpose in this post is just to bring you these figures so that you'll a realistic picture of what is happening (something you absolutely cannot get that from hysterical news reports).

Bob Woodward is Misleading the American people

I just noticed an editorial by Bob Woodward in today's Washington Post (via RealClearPolitics) that is preposterously entitled "Secret Reports Dispute White House Optimism." In that editorial, he says:

Two days later, the intelligence division of the Joint Chiefs of Staff circulated a secret intelligence assessment to the White House that contradicted the president's forecast.

Instead of a "long retreat," the report forecast a more violent 2007: "Insurgents and terrorists retain the resources and capabilities to sustain and even increase current level of violence through the next year."

A graph included in the assessment measured attacks from May 2003 to May 2006. It showed some significant dips, but the current number of attacks against U.S.-led coalition forces and Iraqi authorities was as high as it had ever been -- exceeding 3,500 a month. [In July the number would be over 4,500.]

Bob Woodard wants you to believe that all of this is secret information, and he wants to present himself as the one who is boldly revealing this news to the American people. But he is living the past, pathetically trying to resurrect his glory days. It's almost sad.

As I already noted here and here, all you need to do is go to the Iraq Index. Check pages 22 and 23 to see that, indeed, the number of attacks against coalition troops is as high as it has ever been. According to one graph, it's 90 per day (about 2800 per month):


According to another, it's now 800 per week (about 3200 per month):


Yet another graph shows that IED attacks are high as well:



Woodward cites an even higher number (perhaps because it includes additional attacks on Iraqi authorities), but the key point -- that attacks are at an all-time high -- is not secret. It's been right out there for everyone to see. Woodward knows that most people don't read the Iraq Index on a regular basis, so they won't realize that he's not boldly revealing secret information. But I do, and I know.

And I also know that it's silly to focus solely on that one single statistic (i.e., the number of attacks) when other statistical indicators show a different trend. In particular, the increasing number of attacks appear to be decreasingly effective (all data here):



So it's easy to see why some are optimistic about our fight against the insurgency while others are pessimistic. Optimism and pessimism are affected by the evident trends, and the trends here do not agree. Attacks are up (warranting pessimism), but, more importantly, casualties are down (warranting optimism). That would appear to suggest a weakening insurgency despite an increased number of increasingly ineffective attacks.

Still, I can see how some might be pessimistic. But for Woodward to imply that Bush is hiding the secret details of a vastly worsening insurgency while faking optimism about what he secretly knows to be a worsening situation is to play a silly little game designed to increase the number of people who will watch his interview on 60 minutes tonight and buy his book tomorrow. It's fine for him to do that, but I agree with John Dickerson, who wrote an article for Slate entitled:

Bush vs. Woodward
Only one reputation can survive.

Dickerson got the title right, but he picked the wrong survivor. Woodward's reputation is the one that will not survive this ridiculous little stunt.