Fosberry's Blog

An Inconvenient Obligation

2009-01-26T10:51:48Z

President Obama's inauguration speech was full of forward looking bipartisanship . We should not let partisan sniping stand in the way of doing the work necessary to fix our economy, improve our nation's health care, and develop alternative energy sources.

Yet there is one thing which Obama still must do, even though he will be accused of partisanship by doing so: he must thoroughly investigate Bush administration, and, where evidence of criminal action is strong, prosecute those responsible.
]]> The human rights abuses of the Bush administration must be thoroughly investigated, not merely because they were wrong, but because we have an obligation to do so under treaties our nation voluntarily signed, including the U.N. Convention Against Torture.

Article 12 of that treaty says that "Each State Party shall ensure that its competent authorities proceed to a prompt and impartial investigation, wherever there is reasonable ground to believe that an act of torture has been committee in any territory under its jurisdiction."

In the aftermath of Susan Crawford's admission that military personnel tortured Mohammed al-Qahtani, there is certainly "reasonable ground" to believe torture occurred. And, while the administration's choice of putting a detention center on Guantanamo intentionally left it outside the jursidiction of U.S. law, it's hard to argue that Guantanamo is not territory under American control.

This treaty allows no exceptions for "national security" or for dealing with "terrorists" or "enemy combatants" who are not deemed prisoners of war. From article 2:
"No exceptional circumstances whatsoever, whether a state of war or a threat of war, internal political instability or any other public emergency , may be invoked as a justification of torture". (emphasis added)

If for some bizzare reason the Obama administration feels compelled to continue torturing, it may denounce the treaty, which the United States signed in 1988, under Ronald Reagan, and ratified in 1994, under Bill Clinton. Article 31 allows a nation to "denounce" the convention, and one year after if formally does so, that nation is no longer bound by the convention's restrictions. But article 31 paragraph 2 reads:
"Such a denunciation shall not have the effect of releasing the State Party from its obligations under this Convention in regard to any act or omission which occurs prior to the date at which the denunciation becomes effective".

In plain English, even if we withdraw from the convention (itself a horrible public relations move), we're still obligated to investigate torture, and, where evidence of it exists, we must prosecute.

This isn't some political witch hunt of partisan Democrats exacting "vengeance" on hated Republicans. It is an obligation under a treaty freely signed by a Republican president, and later ratified by a Democratic Congress. And it's the moral thing to do, too.

McCain 2008: Good and Bad for the GOP

2009-01-05T11:34:21Z

My friends, I come to praise Senator McCain, not to bury him. Despite lame speeches in front of green backgrounds, his ostrich-like insistence that the fundamentals of our economy were sound, and and bestowing the gift of Sarah Palin's VP nomination on thankful political satirists, I still think McCain was the strongest candidate the Republicans could have run in 2008. Yet running their strongest candidate was a bad strategic choice for the GOP.
]]> Certainly McCain's campaign made its share of mistakes, but but he still won the votes of nearly 60 million Americans, more than all but one Republican candidate ever (W. got 62 million votes in his 2004 reelection), and he still got over 45% of the popular vote. What Republican could have run a stronger campaign? Mitt Romney would have been far easier to tie to President Bush and the Republicans than the maverick McCain. Branding is a powerful factor in popular perception, and McCain's image as a moderate Republican was vital to his (relative) electoral success. Mike Huckabee lacked the support of fiscal conservatives , who recoiled against his economic populism. Rudy Giuliani or Fred Thompson might have been appealing on paper, but neither showed the inclination to, well, actually campaign for the office.

And recall that any nominee would have made his share of mistakes (Romney's saying his sons had "more important" things to do than serve in Iraq, for instance) that would have blunted their appeal to a broader electorate. And, while she became a national laughing stock to most, Sarah Palin did solve McCain's core problem of an unenthusiastic right-wing base: after choosing her, that base came in strongly for the Arizona Senator. So I don't think the Republicans could have run a stronger candidate than McCain.

Yet I also think McCain's candidacy was a net negative for the GOP. Because of his "maverick" image, and his demonstrable mistakes, many on the right are deluding themselves into thinking that the problem was that the Republicans didn't run a "true conservative". They look at McCain's mistakes and console themselves that the problem was the candidate and the ineptness of his campaign, not the message. Ironically, the Republicans' long-term interests would have been better served by a crushing defeat of, say, Romney, running as a "pure" Reagan conservative than by McCain's not really close loss. Had they seen more clearly the country's rejection of the message, they would be forced to do more serious self-examination of what, other than reflexive opposition to "Liberalism", the party stands for.

The Republicans are becoming a regional party, genuinely strong only in the south and the sparsely populated plains states. Its elected officials, the main pool for future potential national candidates, is growing more conservative as the country is shifting more towards the left. And the growing homogeneity of the Republican party ("liberal" Republicans, or even pro-choice ones, have been purged, resulting in the party's death in the northeast) helps ensure that any future national candidates will stand far to the right of the American people.

McCain, despite his maverick image, was much more conservative than most Americans thought, but the growing consensus among Republicans is that it was his centrism and moderation that hurt him. And because he wasn't seen as a true believer conservative, the conservatives who are now a firm majority within the party don't see his defeat as a defeat of conservatism itself. They instead see it as a rejection of a weak candidate who ran a poor campaign (e.g. why didn't they push Rev. Wright as a campaign issue?). So the reassessment that the Republicans need has been postponed, and the party drifts further to the right on social issues in a country that is moving to the left.

This is good for Obama's political fortunes. I ascribe to the Clintonian credo that "it's the economy, stupid" that drives presidential elections. Obama is clearly inheriting an economy free-falling into recession, if not depression, so if things recover for 2012, he should coast to a landslide reelection. But even if the slump persists into 2011 or early 2012, Obama may not bear all the blame. Had a more pure Republican candidate (say, Romney or someone like Sarah Palin) lost in a decisive rout in 2008, Republicans might more easily see that on many issues, they're becoming increasingly out of step with more of America.

It is of course far too early to write the obituary of the Republican party, but their initial reaction to McCain's defeat shows more delusion than understanding of why Obama won.

Proposition 8 Not the Final Answer

2008-11-13T09:34:27Z

Proponents of same-sex marriage were understandably saddened by the passage of Proposition 8 in California, but there is good news elsewhere. Yesterday saw the first same-sex marriage performed in Connecticut, following the state supreme court's decision in October overturning a law banning same-sex marriage.
]]>
Yet the Connecticut decision is important also, even though it directly affects a much smaller population. Prejudices fade away slowly over time, and I am confident that Proposition 8 will eventually be overturned, and that marriage equality will ultimately prevail in the United States. And while progress may have taken a large step back in California on election day, it moved forward in Connecticut yesterday, and for that we should be thankful.

The Nader Effect: Keeping Missouri Red?

2008-11-06T19:34:15Z

Ralph Nader won nearly 100,000 votes in Florida in 2000, far more than the 547 vote margin by which George W. Bush was proclaimed the winner over Al Gore, and he took quite a bit of heat for siphoning off voters who likely otherwise would have gone to Gore.

It's arguable that Nader's presence on the ballot again cost a Democratic candidate a state, this time Missouri. According to the TPM results, Barack Obama trails John McCain by just 5,868 votes, with Nader polling 17,769. Had Nader not been on the ballot, if just a third of his supporters voted for Obama instead, the Illinois senator would be ahead in Missouri as well.

This year no one state could tip the election, so this isn't nearly as visible an issue. But it again highlights the effects of the electoral college, and it is a reminder of the impact even minor third-party candidates can have in a close election.

Final State Polling Simulation

2008-11-02T14:43:42Z

Each week I've been taking data from Andrew Tanenbaum's www.electoral-vote.com and running Monte Carlo simulations of the election. I do two runs, one using a 4% margin of error (common for most state polls), and another using a margin of error taken from a linear regression of state polling results in 2004 compared to the final election results. This week, that value is a 10% margin of error.

This week's results:
4% margin of error
Obama wins 100.0%, averages 359.2 EV (low 311, median 363, high 393)
McCain wins 0.0%, averages 178.8 EV (low 145, median 175, high 227)
No electoral ties

10.0% margin of error
Obama wins 100.0%, averages 354.6 EV (low 277, median 356, high 442)
McCain wins 0.0%, averages 183.4 EV (low 96, median 182, high 261)
No electoral ties

These results are not materially different from either of the past two weeks.

This week saw nearly 200 new state polls, with 14 in Pennsylvania, 11 in Virginia, 12 in Ohio, 10 in North Carolina, 11 in New Hampshire, and 12 in Florida. There is probably some tightening in Pennsylvania, with the three most recent polls showing Obama leads of 4, 6, and 7 points, while earlier surveys released this week showed Obama leads as high as 14 points. Obama is still clearly ahead here, and McCain needs more movement to make this state competitive.

Florida is still tight, but might be trending more towards Obama: he leads in 11 of the 12 new polls, and is tied in the other (Zogby's only poll of Florida this year), but his largest lead is only 5 points. Ohio looks better for Obama right now, as he led in all 12 polls this week, from 3 to 16(!) points - the 16 point (an Ohio University poll) is pretty clearly an outlier, as no other pollster showed Obama up by more than 9. But McCain hasn't led an Ohio poll since mid-October.

North Carolina is also tight, but still leaning Obama. He led in 7 of 10 polls by between 2 and 7 points, with two ties, and one McCain lead of 1 point. Virginia and New Hamphsire look more solid for Obama, as he led in all the new polls, at times by double-digits. McCain hasn't led a Virginia poll since early October, and he's not led in New Hampshire since late September (while Obama has been over 50% in each October poll of the Granite State).

Indiana genuinely looks like a toss-up now, with a slight McCain lean. Of 7 new polls, McCain led in 3 (by 3, 6, and 2), Obama led in 2 (by 1 and 2), and two others were tied. Missouri might even be closer, with 2 ties, three with Obama leads, and three with McCain leads. Obama led by at least 4 in all 8 new Colorado polls, and also by 4 or more in all 6 Nevada polls.

Arizona has seen a flurry of new polling activity, with 8 new polls this week after just 16 earlier in the year. McCain has still led in every poll this year, but his lead has shrunk this week. One was as close as 1 point, and two others were at 2 points, while McCain's biggest leads were 8, 7, and 5. Montana had 4 new polls, 3 showing McCain ahead by 4, and the fourth showing McCain up 3. North Dakota's only poll had McCain up by just 1, while five new Gerogia polls showed McCain leads of between 1 and 6.

The key point is that with the possible exception of Pennsylvania, none of the Kerry states are remotely close. But among Bush states, Obama leads comfortably in Iowa and New Mexico, with Colorado, Virginia, and Ohio looking quite likely to go blue. Even North Carolina and Florida are tilting to Obama, and he could win other traditionally red states like Indiana, Montana, or even McCain's home state of Arizona. Barack Obama is overwhelmingly likely to be elected our next president.

Penultimate Polling Simulation

2008-10-25T14:56:59Z

Each week I've been taking polling data from Votemaster Andrew Tanenbaum's www.electtoral-vote.com and running Monte Carlo simulations of the general election. I do two different runs, one using the 4% margin of error that is common to most state polls, and another using a higher margin of error that I derived from a linear regression of state polling data in 2004 compared to the final election results.

This is the first week's results:
4% margin of error:
Obama wins 100.0%, averages 368.8 EV (low 317, median 375, high 396)
McCain wins 0.0%, averages 169.2 EV (low 142, median 163, high 221)
No electoral ties

10.2% margin of error:
Obama wins 100.0%, averages 357.1 EV (low 277, median 358, high 435)
McCain wins 0.0%, averages 180.9 EV (low 103, median 180, high 261)
No electoral ties

Last week, McCain won just one of the 20,000 simulations, but this week Obama won all of them, and he increased his average and median electoral vote totals. Obviously this is the most favorable the state polling has been to Obama all year, and it comes at a good time. Most of the new state polls have been favorable to Obama. Indiana's two new polls gave Obama leads of 5 an 10(!) points, where previously the state had been tied or had McCain leading. It is likely an outlier, but one of two new Georgia polls actually showed Obama ahead by 1, with the other giving McCain a 5 point lead, still a tighter race than we had seen before. Georgia now looks like North Carolina did in August. Speaking of the North Carolina, it had 8 new polls released this week, with McCain leading by 2 in one, another tied, and Obama ahead in the other six with margins ranging from 1-4 points. The Tar Heel state is still quite close, but Obama appears to have the upper hand. Neighboring Virginia looks more likely to vote for Obama, as he leads in all 5 new polls, with two different surveys showing a 10 point lead, but another has the lead at just 1.

Florida is still close, but leaning Obama, with McCain having a 1 point lead in 2 new polls this week, but Obama leading by 1, 5, and 7 in the other three polls.

Obama got strong polling in the midwest. McCain led in 2 of 7 new Ohio polls, by 1 and 2 points, but Obama led the other 5 by wider margins, including the first two polls giving Obama leads of more than 10 points in the Buckeye state. Indiana
Perhaps Pennsylvania is tightening a bit, but Obama's lead in the 5 new polls this week were still all at least 10 points. The two new Indiana polls both have Obama ahead, by 4 and by 10 points.

Montana joined North Dakota as a formerly solid red western state that might turn blue, as two new polls split, with each candidate getting a 4 point lead in one. Even Alaska may be tightening some. While Obama won't win, McCain's lead has shrunk from a high 29 points just after the GOP convention to just 11 points in a new poll this week.

Maine, where McCain was hoping to make a play for one of the congressional districts, now looks solidly Obama, with three new polls giving him leads of at least 15. One of the few relative bits of good news for McCain was in New Hampshire, where Obama's leads in the two new polls are just 4 and 7 points, closer than the double-digit leads seen in early October. Missouri is still tight, with the two new polls differing on the leader, but the Obama lead is 5, while the McCain lead is just 1, suggesting Obama may be slightly favored there, also.

Pennsylvania looks solidly blue, with the closest of 5 new polls showing an 8 point Obama lead. West Virginia looks like it will stay with the GOP, as McCain had leads of between 4 and 9 points in the 4 new polls, and Wyoming looks safe.

Barring some entirely unexpected news to shake up the race, Barack Obama will be elected as our forty third president.

Colin Powell's "Really Right Answer"

2008-10-20T12:56:04Z

What struck me most from Colin Powell's statements in endorsing Barack Obama on Meet the Press yesterday was this quote:

Well, the correct answer is, he is not a Muslim; he's a Christian. He's always been a Christian.
But the really right answer is, what if he is? Is there something wrong with being a Muslim in this country? The answer's no, that's not America.
Is there something wrong with some seven-year-old Muslim-American kid believing that he or she could be president?

Yes, McCain corrected the woman who claimed Obama was an Arab, and he pointed out he was a Christian. But whether Obama is a Christian, Muslim, Jew, Hindu, Buddhist, or atheist should be irrelevant to our political discourse.

Powell gets that, and it was a major factor in his endorsement.

He also closed with a moving anecdote, describing a photo of a mother grieving at her son's grave in Arlington Cemetery. At the top of the headstone was neither a cross nor a Star of David, but the crescent moon and star representing Islam. That underscores that America is not just one religion, but many, and it is that diversity that helps make America a great nation. As says on our money, e pluribus unum - out of many, one.

Antepenultimate Polling Simulation

2008-10-18T13:18:37Z

Each week I've been taking polling data from Votemaster Andrew Tanenbaum's site www.electoral-vote.com and running simulations of the general election. I do two versions, one using the 4% margin of error common to state polls, and another using a margin of error derived from a regression of 2004 polling data against the actual results. The former is a snapshot of where we are now, and the latter allows for more movement in states before the election.

This week's numbers are even better for Obama than last week's.

4% Margin of Error:
Obama wins 100.0%, averages 360.9 EV (low 307, median 358, high 387)
McCain wins 0.0%, averages 177.1 EV (low 151, median 180, high 231)
No electoral ties.

10.3% Margin of Error:
Obama wins 99.99%, averages 346.6 EV (low 268, median 348, high 419)
McCain wins 0.01%, averages EV (low 119, median 190, high 270)
No electoral ties.

Yes, out of 20,000 trials with two different margins of error, McCain won exactly once. Obviously the one trial that resulted in a McCain victory has no statistical significance, and if I reran the simulation, I'd likely get different results - perhaps a few more McCain wins, or none. But it's safe to say that based on current polling, McCain effectively has no chance to win.

Just for fun, I looked at the one case McCain won in more detail. It was mostly what we'd expect. He takes must-win states Florida, Ohio, Virginia, North Carolina, Missouri, and Indiana. He also wins Colorado, New Mexico, and Maine(!), while Obama wins the other Kerry states plus Iowa, Nevada, West Virginia(!), and North Dakota(!!). This highlights that my simulation assumes each state is a completely independent trial, as it's almost inconceivable that Obama might lose Maine but win North Dakota and West Virginia. I should also note that I'm not attempting to allocate Maine's or Nebraska's congressional districts separately: I'm assuming these states are winner-take-all like the rest.

The new polls are again almost all favoring Obama. He's shown consistent leads in Colorado (4 new polls, ranging from +4 to +9) and Virginia (3 polls, +3, +6, and +10), and Pennsylvania (4 polls, +8 to +15). Florida had 6 new polls, one showing McCain up 2, but the other 5 all gave Obama a lead from between 2 and 5 points. Similarly Ohio had 5 new polls, McCain leading by 2 in one, one tied, and 3 with Obama leads of 2, 5, and 5. North Carolina is still a toss-up, with 3 new polls evenly split between a tie and 2 point leads each way. Missouri may be breaking towards Obama (4 polls, one with McCain up 1, the others with Obama leads of 3, 6, and 8).

Obama is right to be worried about overconfidence, because this polling data sure looks good.

Obama Extends Polling Lead

2008-10-11T17:31:11Z

Each week since the middle of the primary season I've been running Monte Carlo simulations of the general election using polling data from Votemaster Andrew Tanenbaum's www.electoral-vote.com.
This week's polling data shows has been the most favorable for Barack Obama all year. Without some significant change in the electoral landscape, John McCain simply can't win.

I'm still doing two simulations, of 10,000 trials each, but with a different margin of error. First is a 4% margin of error common to most state polls, a run which reflects what would happen if the election were held now. The other uses a margin of error I computed using a linear regression of 2004 polling data errors compared to the final result, with time to the election as the independent variable. That now uses a 10.5% margin of error.

4% Margin of Error:
Obama wins 100.0%, averages 353.5 EV (low 296, median 354, high 384)
McCain wins 0.0%, averages 184.5 EV (low 154, median 184, high 242)
No electoral ties.

10.5% Margin of Error:
Obama wins 99.94%, averages 340.6 EV (low 247, median 343, high 403)
McCain wins 0.06%, averages 197.4 EV (low 135, median 195, high 291)
No electoral ties.

The 10.5% margin of error does overstate Obama's chances of winning (since each state trial is independent; it's more likely that if McCain gets momentum, he'd gain in many states at the same time, as Obama has over the past two weeks), but not necessarily by that much. Nate Silver now estimates Obama has a better than 90% chance of winning. Obama is clearly in a dominant position.

Almost all of the recently toss-up states have moved strongly towards Obama.
The two latest Colorado polls both gave Obama a 6 point lead. Obama led in all four new Florida polls, with margins between 2 and 7 points.
Likewise Minnesota's five new polls showed Obama with leads, ranging from 1 to 18(!) points. Obama led by 2 and 7 in the two new Nevada polls, and he led in the three new New Hampshire polls by 9, 13, and 8.
Ohio had 7 new polls, with McCain leading in just one (by one point), and Obama ahead in the other 6, by 3-6 points. Obama's leads in the 5 new Pennsylvania polls were all 10 points or more. Even traditionally red Virginia had two new polls giving Obama double-digit leads (the third still had him ahead by 2).

North Caolina is the lone exception, perhaps still slightly leaning to McCain, with three new polls splitting Obama +1, tie, and McCain +3.

Other previously safe Republican states are now perhaps in play. The two new polls in Missouri both showed a 3 point lead, but differed on who was ahead. The first new poll of West Virginia in over two weeks showed Obama ahead by 8. Three new polls in Georgia showed McCain ahead by 7, 9, and just 3. A new poll of Montana saw McCain's lead shrink to 5.

As the financial crisis has dominated the news, Obama has pulled further ahead. The McCain campaign seems disorganized and unfocused, unable to settle on a consistent message. Its handling of the Ayers association is a good case in point - at times they are trying to paint Obama as unacceptable and perhaps a terrorist sympathizer, but McCain backed away from such direct accusations in the town hall debate. One night a campaign spokesman says it's not an issue, the next day another raises it. McCain did the right thing by calling out some supporters who got out of line with outrageous statements, like calling Obama a terrorist, but much of his campaign is essentially pushing that very idea.

It is a campaign that knows it is losing, but has run out of ideas for yet another tactical diversion, and so it may now be resigning itself to defeat.

Putting country first? Or simply lying?

2008-10-09T12:56:01Z

In Tuesday night's debate, Senator McCain said:

I'll get Osama bin Laden, my friends. I'll get him. I know how to get him.

I'll get him no matter what and I know how to do it. But I'm not going to telegraph my punches, which is what Sen. Obama did.

If McCain is to be believed, then why, in the 7 years since 9/11, has he not told the Bush administration how to get Osama bin Laden? Does he possess some superpower that will only enable him to get bin Laden if he's serving as
president?

One of the following must be true:
1. McCain was simply lying about knowing how to get bin Laden.
2. He doesn't put country first, and won't say how to get bin Laden until after he's elected.
3. He believes his own election is more important to the country than capturing the perpetrator of the deadliest terrorist attack on American soil, so letting bin Laden stay free actually is putting his country first, at least in McCain's
mind.

My money is on choice #1, but none of the three options speak well of McCain.

Obama Takes Commanding State Poll Lead

2008-10-04T18:19:38Z

Each week I've been taking polling data from Andrew Tanenbaum's www.electoral-vote.com and running Monte Carlo simulations of the election. I use two different margins of error, a 4% margin that is common to most state polls, and another I computed by running a regression of polling errors in 2004 data compared to the final election results. The 4% margin of error is a snapshot of who would win if the election were held today. The other simulation now uses a 10.6% margin of error, which was an attempt to allow movement before the election.

Obama wins almost all the time in both simulations this week/
4% Margin of Error:
Obama wins 99.99%, averages 340.9 EV (low 269, median 344, high 375)
McCain wins 0.00%, averages 197.1 EV (low 139, median 194, high 269)
Electoral tie 0.01%

10.6% Margin of Error:
Obama wins 97.86%, averages 322.6 EV (low 230, median 323, high 419)
McCain wins 1.99%, averages 205.4 EV (low 119, median 215, high 308)
Electoral tie 0.15%

This is easily Obama's most commanding lead of the year, in both simulations. The 10.6% margin of error clearly overstates Obama's chances of winning, because it is assuming each state varies independently. If McCain makes a comeback, he's likely to do so in most states at the same time.

Needless to say, this week's state polling data were quite favorable to Obama. He's probably moved ahead in Florida (leading in 4 of 6 polls, with one tied, and the other giving McCain just a 1 point lead), and is polling a little stronger than McCain in Virginia (leading in 3 of 5 polls), Ohio (4 new polls split, with each candidate leading 2, but Obaa's leads are larger). North Carolina had two new polls, with each candidate showing a 3 point lead, so that state may well be a toss-up right now (McCain had led by double-digits immediately after the GOP convention). McCain gave up on Michigan, but Pennsylvania has shifted more solidly blue also, where the closest of 4 new polls was a 5 point lead, while another poll showed Obama ahead by double-digits.

And while it didn't affect the 4% margin of error simulations, McCain's leads in some traditionally red states shrunk also, with Georgia, Mississippi, and West Virginia all showing McCain leads under 10%. In the simulations, McCain was more likely to lose West Virginia than he was to win Pennsylvania. Yes, Obama now has more very small leads, but adding up the states where he's ahead now puts Obama at almost 340 electoral votes, so he can afford to lose several big, close states and still win.

It's just a month after McCain was rated as more likely to win than Obama, and there's still one month to go, so it's possible that things could swing back the other way. Only four states(West Virginia, Wisconsin, Michigan, and Maine) which last showed a lead of under 10 points didn't have a new public poll this week. McCain's decision to pull out of Michigan implies his internal polling was not favorable.

The state polls don't yet include reactions to Thursday's vice presidential debate, but it doesn't seem like it was the game-changer McCain needs to make this competitive. Obama is now in a very strong position, and while these numbers overstate the chances, he is now highly likely to be our next president.

Late September Election Simulations

2008-09-27T20:24:06Z

I've run another set of simulations of the election using data from Votemaster Andrew Tanenbaum's www.electoral-vote.com. State polling data is now coming fast and furious. The least recently polled state, Louisiana, had an average poll date of September 12, and just 16 states haven't been polled within the past week. So most of these results incorporate voter reactions to the Lehman Brothers bankruptcy, and the Federal Reserve's takeover of AIG, with over half coming after Henry Paulson and Ben Bernanke proposed their $700 bailout of distressed mortgage assets.

I'm still running a 4% margin of error snapshot, which represents what would happen if the election were held now, and also a 10.75% margin of error, which is about how accurate the polls were in 2004 this far out. The bounce John McCain got in the aftermath of selecting Sarah Palin and the GOP convention has now fully dissipated, with Obama posting about as strong numbers as he had all year.

4% Margin of Error:
Obama wins 96.7%, averages 299.5 EV (low 243, median 297, high 370)
McCain wins 1.6%, averages 238.5 EV (low 168, median 241, high 295)
Electoral tie 1.7%

10.75% Margin of Error:
Obama wins 87.0%, averages 298 EV (low 203, median 298, high 390)
McCain wins 12.2%, averages 240 EV (low 148, median 240, high 335)
Electoral tie 0.8%

Obama has a clear lead, but as we saw with the GOP bounce, things can change quickly. It is interesting that McCain ties about as often as he wins outright in the 4% margin of error case. The most common tie scenario there remains Obama winning all the Kerry states except New Hampshire, plus Colorado, Iowa, and New Mexico, with McCain winning the rest of the Bush 2004 states. Even the wider margin of error doesn't help McCain much any more, as recent polling has stengthened Obama's once tenuous leads in Michigan and New Mexico, and states the GOP convention seemed to take out of play, like FL, NC, and MO, are again competitive.

Almost all the changes this week were good for Obama. Obama had leads of 4, 4, 3, and 9 in Colorado, with one outlying poll (from ARG) showing McCain up 3. Florida, which looked like it might be drifting out of reach, is now basically a tossup again, as 4 new polls gave McCain leads of 1, 5, 2, and 1, while two other polls (ARG and Mason Dixon) gave Obama leads of 1 and 2. Michigan had 8 new polls, with one showing a tie, one giving McCain a 3 point lead, but the other 6 giving Obama leads of 4, 5, 7, 8, 10, and 13.

8 new polls in Pennsylvania all showed Obama head, by between 2 and 9 points. Ohio remains leaning to McCain, where one poll showed a tie, but two McCain 4 and 1 point leads. Like the tar heel state, however, the more recent poll was weaker for McCain. Virginia had 7 polls, with Obama leading in 4 and McCain in 3.

It seems that the continuing financial crisis isn't helping McCain. Rasmussen released two polls in each of four different states, and in all four, the more recent poll was better for Obama. Obama increased his Pennsylvania lead from 3 to 4, and he cut McCain's Ohio lead from 4 to 1. But he did even better on Tobacco Road, turning a 3 point North Carolina deficit into a 2 point lead, and taking a 5 point lead in Virginia after trailing be 2 before.

The news wasn't all bad for McCain - after looking like it might be competitive, Montana now looks safely red, with two new polls giving McCain 9 and 13 point leads.

The state polling appears to lag the daily tracking polls a bit, but the recent momentum has been entirely in Obama's direction. Presumably the McCain campaign knows all this (and arguably has even better data), so I have little doubt that his attempt to "suspend" his campaign was an intentional campaign tactic designed to try to shake things up again. The current narrative wasn't working, and he needed to try something to change the game. The Palin selection had worked well tacticially (although it perhaps now may be turning into a net negative), so McCain hoped to do some other "maverick" like thing to shake things up. It looks like that hasn't worked, and last night's debate, which both sides had reasons to be happy about, also likely didn't provide McCain a significant boost.

The clock is ticking for the McCain campaign, and Obama is still keeping his cool, and his lead.

McCain's Solution to the Financial Crisis?

2008-09-24T21:19:16Z

He intends to sit all the subprime borrowers and lenders down in a room and tell them, "Stop the bullsh*t!".

The Fed Chariman has no Clothes

2008-09-24T14:24:53Z

In arguing for the $700 billion bailout plan, Federal Reserve chairman Ben Bernanke says that the government should pay "hold to maturity" prices for assets, which are currently notably higher than the market prices private investors are willing to pay.

Bernanke's stated belief is that the government can pay more for these assets, but because the market is underpricing them now, it can effectively subsidize financial firms (by paying more than market prices) without risking taxpayer money (because if held long enough, the assets will be worth more than he's paying).

Markets can be irrational, both in bidding up assets and in bidding them down, so it is possible that Bernanke could be right with his guess. But by talking of "hold to maturity prices", Bernanke is speaking of something that does not exist. The cash flows from these assets are highly dependent on default rates and recovery rates in default situations, and those are highly dependent on both real estate prices and the strength of the overall economy: if your house is worth less, or if you lose your job, you're much more likely to default. And the more people who are defaulting, the more house prices fall as foreclosed properties hit the market, which further reduces the value recovered when defaults occur.

Bernanke is pretending that precision exists, and that there are assets that, if only held to maturity, will pay at par. The reason current market prices, where they exist, are low is precisely because of the risk: people don't know how these things will perform in future scenarios. You can use models to run simulations, and to chart out what you believe to be optimistic or pessimistic scenarios, but it's still guesswork.

This is why the price for entry to any bailout must be some serious equity in the firms that participate. I agree with Bernanke that to recapitalize the banks, the Treasury would need to pay more than current fire sale prices, but to lower the risk to the taxpayer, and to keep in place some sort of moral hazard for troubled financial firms, the government needs to get more than troubled assets for its recapitalization. Sure, more solvent banks might balk at dilution, in which case they can absorb losses on their own. Requiring an equity stake as a condition for a deal will ensure that the bailout funds only go to those institutions that are in the most need, and also that, should those institutions survive, the taxpayer gets some participation in their recovery.

Weekly Election Poll Simulation

2008-09-21T19:44:55Z

There have been lots of new state polls this week, and I've rerun my simulations of the general election, one assuming a 4% margin of error for the state polls, and another using a 10.9% margin of error. The first is a snapshot of what might happen if the election were today; the latter uses a margin of error calculated from a linear regression of errors 2004 state polls had compared to the final results, assuming the error would fall as we came closer to the election.

This week's results:
4% Margin of Error
Obama wins 51.6%, averages 270.1 EV (low 212, median 272, high 338)
McCain wins 42.4%, averages 267.9 EV (low 200, median 266, high 326)
Electoral tie 6.1%

10.9% Margin of Error
Obama wins 53.8%, averages 272.5 EV (low 160, median 273, high 362)
McCain wins 44.5%, averages 265.5 EV (low 176, median 265, high 378)
Electoral tie 1.7%

The most common tie scenario involves Obama winning all the Kerry states except New Hampshire, and Obama also winning New Mexico and Colorado.

The polling is now recent for almost the entire country. Only Arkansas and Massachusetts have not been polled since the end of the GOP convention.

With so many polls, it's possible to see some rather large disagreements. For example, two polls last week gave Obama a double-digit lead in Iowa, but a third had the race tied. Three polls of North Carolina showed McCain ahead by 11, 1, and 6. There were 9 new polls in Ohio, ranging from Obama up by 2 to McCain up by 6. Two polls showed a tie in Pennsylvania, but a third showed Obama ahead by 5. Three Florida polls showed a tied race, but three other polls gave McCain leads of 1, 5, and 6. Colorado had 4 new polls, with McCain leading by 2 in 2 polls, Obama ahead by 1 in another, and Obama ahead by 10 in the last.

Much of these fluctuations are probably simply sampling error, but pollsters also can differ in how they weight their raw data.

Things are tight right now, and it will be interesting to see how the week's disruptions on Wall Street affect the next round of polling.