O'Neil Perspective

What does the testimony of a Presidential nominee to a nonpolitical federal statistical agency tell you about the strategic approach of the Obama administration? If you observe how it fits a pattern, it tells you quite a bit.

Click here for the remainder of the article: http://www.huffingtonpost.com/michael-j-o/census-pick-illustrates-b_b_215847.html

Michael J. O'Neil, PhD

Michael J. O'Neil, PhD

My take on the first hundred days of the Obama Presidency, as broadcast this morning on ABC15 TV.

Click here to view.

Cheers,

Michael J. O'Neil, PhD

(Published in the Phoenix Business Journal - March 27th, 2009)

Everyone, it seems, is furious about the AIG bonuses.  Almost everyone (myself included) shared the same initial reaction: HOW DARE THEY take millions in bonuses out of the billions in public funds they received!    Especially because it was the risky calls they made that produced the need for public funds--calls that resulted in billions of dollars of losses.  IT IS OUR MONEY!
 
But Anger is usually a lousy basis for sound public policy.  Politicians are jumping over one another to try to figure out how to tax back every penny the greedy bastards stole from the public, full well knowing that there is a better than even chance of the courts striking such laws down as unconstitutional. (If you don’t understand why, go to Wikipedia and search for either “ex post facto” or “bill of attainer”).  Since the politicians are aware of this, their actions are more designed to show action than to effect it.
 
More importantly, anger-based policy actions almost always neglect larger, more structural, problems.  The issues involved are much larger than AIG specifically, or even bailed-out companies.  The entire basis of compensation in much of the upper reaches of the corporate world has come to bear little relation to any known reality.  It was not always the case, but corporate C-level compensation in America has grown to be out of line with the rest of the world and it is largely out of the reach of even company stockholders. The financial world is a case in point, but only a manifestation of a larger trend.
 
AIG employees who traded in credit default obligations (CDOs) and similar products dealt in very large denominations. Huge sums were made by AIG, and the traders were paid commissions on these profits.  And no one objected, since these geniuses were making much larger sums for AIG.
 
That worked until sometime in 2007 when the underlying structure of the market went to hell. Now the same geniuses who had made billions for AIG were losing billions for the company.  How could these geniuses turn into morons so quickly?
 
Of course, they were never that smart, nor that dumb.  Almost anyone with a modicum of numerical competence could have made fortunes doing what they were doing through 2006 and the early months of 2007.  And, if they stayed in the same business, even the best and the brightest of them could not have prevented massive losses from sometime in 2007 forward.
 
And insiders were savvy enough to recognize this.  They got AIG to guarantee them the same level of bonuses in 2008 and 2009 that they had received in 2007.  Everyone knew the bottom was falling out of the business. So they covered themselves.
 
But part of the explanation for the attitudes of those involved may be found in human nature. When things go well, we all tend to attribute it to our intelligence and hard work.  When they go badly, we all blame bad luck or forces beyond our control. These traders were no different.
 
If we focus only on getting the ill-begotten bonuses back, we miss the larger point: the compensation systems were irrational from the beginning. Traders were compensated for the sums they “made” for AIG, but not penalized for the long-term risk they also created in the process. And there is a point in all of this about corporate compensation systems generally. Look no further than the auto industry: for decades these behemoths were run by their own brand of geniuses who never did anything other than what had been done before. And they all figured they were pretty smart-and deserving of stratospheric compensation.  But because their genius lay only in doing what had worked for decades, when the world changed, they were the worst possible people to lead their companies into a changed future.
 
And we should not lose sight of the fact that the bonuses involved comprise less than one-tenth of one percent of the sum the federal government has plowed into AIG.  Where is the outcry about the other 99.9+%?
 
 
Michael J. O'Neil is President of O'Neil Associates, a national public opinion and market research firm based in Tempe, Arizona.  O'Neil Associates has conducted more than 1,700 opinion research surveys in the last 30+ years. This  experience has given him ample opportunity to observe that the big guys are not always the smartest. A sociologist by training, he has taught at several universities including the University of Michigan, the University of Illinois, and Northwestern.  He is reachable at oneil@oneilresearch.com or www.oneilresearch.com..

Michael J. O'Neil, PhD

And there is a point beneath the levity.  Invest half a minute to see what it is.  Click here to see it.

Michael J. O'Neil, PhD

I had some comments on the Arizona budget crisis and choices faced by the new Governor, Jan Brewer.  These were broadcast on KNXV/ABC 15 TV yesterday.  Click here to take a peek.

Michael J. O'Neil, PhD

The significance of the Barack Obama’s election hit me especially hard when the results came in from Virginia.  While the Presidential outcome was already evident to me by that time, there was something about the Virginia results that jarred me.  (More)

This past Sunday, I participated in Sunday Squareoff   (NBC, TV 12) with John Kavanagh (chair, House Appropriations Committee) and David Lujan (House Democratic Leader) in a discussion of the Arizona budget. The differences in the positions of these two leaders are striking and clearly reflect the varying perspectives of their respective parties.  Kavanagh (and most Republicans) wishes to cut whatever amount of spending is required to avoid any tax increase, while Lujan (and most Democrats) focus on preserving programs they regard as essential. 
 (More)

 

The map on the right shows the actual results. Projections were made in 49 states; all correctly. The exception: Indiana was projected to be a "lean McCain state"; Obama won it by under 6,000 votes out of over 2.7 million cast. Missouri was adjudged to be an absolute dead heat.  There the margin is under 5,000 out of 2.9 million votes cast; some networks have yet to call the state a week after the election. This is a lot like calling a coin toss and guessing that the coin will land on its side and stay there-and being correct.

Michael J. O'Neil, PhD

Some quick observations on what's next:

http://www.abc15.com/content/news/election/story.aspx?content_id=a58a8f25-6aea-46fb-b36b-e329b386bde8 

(The video is to the right of the page).

Michael O'Neil, PhD 

Contents:

How Networks Do Electoral Projections

How you can know the winner before the network will tell you.

Hour by Hour Guide (poll closing times by state)

Simplified Electoral College Chart (only states in play)

How to use the Chart

Summary Observations

How Networks Do Electoral Projections.  Network projections are made from two sources.  First, exit polls are conducted by an entity contracted by a consortium of all the news networks and several national papers. There is a single entity that does this.  (While it might seem better for each organization to do its own, this is cost-prohibitive.  Think in the neighborhood of $100 million dollars for each separate operation).  The results are compiled and shared with the networks who make independent decisions about how and when to “call” a race.

The poll results will be available at the time the polls close in a given state. If the outcome is clear-cut, an immediate call will be made within minutes of the closing of a state’s polls.  If the race is at all close, it will not be called until enough actual votes are tallied to make the result evident. (If you hear the phrase “Too Close to Call” this is what happened.) But the analysts will not rely on the total raw counts reported, but on random samples of actual votes. In the case of a moderate margin, this will be sufficient to make a projection. But if a state is really close, this may not be sufficient. Then a call would have to await the counting of a much larger number of votes. How many need to be counted depend on how close the race is. In practice, this means that some states are called within a couple of minutes of poll closing while others are called much later.  Indeed, it means that if an immediate call cannot be made on the basis of the exit polls, it will usually be quite some time before a call is made.  How much longer? That will vary widely, depending on the winner’s margin of victory in a given state.

What the networks will not do.  First, they will not call a state before the polls in that state are closed. Second, they will not call the election until they can call states totaling over 270 electoral votes, even though everyone knows that only 17 states are remotely in play (the outcomes of all of the others are beyond any reasonable doubt).

How you can know the winner before the network will tell you.  Download and print the chart by clicking on this link.  It assigns 238 votes to Barack Obama; these are from states whose outcome is beyond question. He needs 32 more for an Electoral College majority. (31 would produce a tie, and he would probably, but not certainly, be elected.)

Should Obama get his 32 Electoral Votes (or if it is clear he will not) the networks will not tell you this outcome until they can project the outcomes from the other states, even if these are the states for which we already know the outcomes.  This is not because they do not know better, but because they are afraid of the criticism they will get if they “call” the election before the polls are closed everywhere. But if you use this chart, you are likely to know the winner before much earlier, possibly hours before the networks will tell you “officially”.  

Why use Obama as the base?  For simplicity: since he has 238 Electoral Votes locked up and is only 32 short, it makes computations simpler.  Rest assured, if Obama does not win, McCain does. If I used McCain as the base, the computations would have been more complicated since his base is much smaller, but the computed outcome would be identical.

What I did. This 238 “safe” Obama electoral vote figure is actually close to the most conservative estimate I could come up with after looking at the polls and several compilations. It does NOT include any “Obama Lean” states; only those rather universally regarded as locked up by him.  The only exception to this I found was that NBC rated Iowa (7 electoral votes) and New Hampshire (4 electoral votes) as “lean Obama” rather than certain Obama. I could find nothing in published polls to support these calls. Each shows double digit Obama leads and neither has had a single poll with a McCain lead in over a month.  I think NBC put them in the Lean category because they each went Republican in one of the two last Presidential elections. (If you want to be really conservative, reduce the Obama number to 227 and add these two states to my list. But it won’t matter, so I kept these states out of the computations for simplicity).

What states are in my 238? While this was not my criterion, my 238 happens to be all of the Kerry 2004 states, plus Iowa (7 electoral votes) minus Pennsylvania (21 electoral votes).  I have already discussed Iowa. PA has clearly been a McCain target; he knows that without it, his winning options are few. And the Republican Party of PA has been running a Rev. Wright commercial starting on Sunday. One wonders what the effect will be of this commercial.  

What else did I exclude from the list?  The “solid McCain” states, of course.  But I also excluded the “lean McCain” states of South Dakota (3), Arkansas (6), West Virginia (5) and Arizona (10).  (Interestingly, some raters   even rate Arizona as a “tossup” state). While Obama might win some of these, I figure if Obama needs any of these to win it means he has lost most all of the tossup states and will most likely lose these as well. If there is an Obama landslide, these lean McCain states could be in play, but they will not determine the outcome. So, again, for simplicity, I omitted them.

There is also one congressional district in NE that Obama could carry for a single electoral vote (McCain will win the rest of the state) and one in ME that McCain could carry (Obama will carry the rest of the state). Neither should matter unless it is very close.

How to Use the Chart.

Ignore all results other than the 13 states indicated on the chart. These are either safe Obama or safe McCain states (and a few lean McCain states excluded for reasons described above). When one of the critical 13 states’ results is determined, enter then number of electoral votes for Obama (either the EV total for the state if he wins, or 0 if McCain wins).

• If this number totals 32 or more, go to bed, Obama is elected even if the networks will not tell you this.  This could be really early in the evening. If this chart is filled out and the numbers you have entered here total less than 31, McCain is elected.

• If the chart is completely filled out and totals something very close to 31 you might want to rip it up and look at those other states and the CDs in ME and NE.  But there is not much chance that will happen. (Remember,  my purpose was to give you something that I am 99% sure will give you the correct outcome and get you to focus on what matters early in the evening without being confused by “calls” of the thirty-some–odd states for which we already know the outcome.  But, if you end up with a number close to 31, you might want to pay attention to the other commentary.
  

Hour by Hour Guide (all times Eastern).

 7pm. I think an Obama victory in either GA or IN would be game-ending. They are a couple of his weakest states on the list. Should he win IN, it will clue that he will probably carry OH and PA as well.  Likewise GA would be a surprise and would probably signal a victory in NC and game over.  But neither of these is highly likely and an Obama loss would not be a major blow to his chances.

Not so Virginia; this is Obama’s strongest state of the three. A VA win for Obama with a PA win later and it is game over.

7:30pm. North Carolina for Obama probably means he carries VA as well and wins the election. Likewise an Ohio victory probably means a PA victory and he wins.

8pm. PA is the closest thing to a “must win” for Obama. With PA there are many win combinations for him. Without it, it gets difficult. FL would nearly cinch an Obama win by itself. Missouri is McCain’s strongest chance of the three.

9pm. If Obama hasn’t cinched by now, NM and CO are likely to be “must wins”.

10pm. Nevada is the only remaining Obama likely state.  If Obama “needs” ND (11pm) and MT (10pm), it probably means he has had such a bad night that these are likely to go for McCain as well.

 Summary Observations

 Think of IN, OH, and PA as in a logical order.  Obama wins IN (the most Republican of the three) and he probably takes them all.  Likewise, McCain wins PA (the most Democratic of the three) and he probably wins all three.

You can think of NC and VA the same way. McCain wins VA indicates a likely NC win. And an Obama win in NC probably signals a VA win.

Of the tossup states, PA CO NM and NV are Obama’s strongest states. Should he win these all and lose everything else he will have 278 electoral votes. Drop CO from this list and we have a 269-269 electoral college tie*—and a scurry to look at those districts in NE and ME. (Right now Obama looks to be an even bet in the NE district; McCain is not so lucky in ME).

The net result: McCain would have to sweep Florida, Georgia, Missouri, Indiana, Ohio AND North Carolina to have a chance.  All of them. And either Pennsylvania or Virginia. And some of the remaining tossup states as well.  While it is possible, it seems unlikely he could win all of these swing states. That is why an Obama victory is probable.

*If there is an Electoral College tie, the House of Representatives elects the President, so Obama wins, since there is a big Democratic majority in the House, right?  Not so fast: it is one vote per STATE. And that Congress has yet to be elected. And that distribution is A LOT closer than you think (a lot of small red states each get a vote equal to that of CA).  Wouldn’t it be ironic to see Obama put over the top in a House vote by a newly elected Democratic delegation FROM ARIZONA?  (And don’t let anyone tell you otherwise: if the election goes to the House, it will be a straight party line vote.).  And no matter who the House selects for President, the Senate will elect Joe Biden Vice President.  McCain/Biden?

“Bradley Effects”, Undecided Voters, Mythical Trends, Likely Voters

Barack Obama goes into Election Day with a substantial lead in both electoral and popular votes. Several reasons have been widely discussed as reasons why McCain could still win the election.  Three lack any convincing evidence.  But two other factors, while unlikely, could still shift the balance toward McCain.

Factors that are NOT working for McCain

No Bradley Effect.  The so-called “Bradley Effect” is based on a reported polling overestimate of the vote for Tom Bradley, a black candidate for Governor of California in 1982.  It has been the source of massive media speculation ever since Obama lost the New Hampshire primary this year. Pundits who knew little about the alleged phenomenon but needed an instant explanation for the anomalous New Hampshire primary results seized upon this largely unsupported theory and have propagated it ever since. I won’t bore you with the details, except to note that the actual concrete evidence for this was scant in 1982 and has been nearly nonexistent since then.

The Undecideds.  This group is ballpark 6% of the electorate. Why won’t they matter much?  First of all, half of them won’t vote. If the other half splits 2-1 for McCain (a fairly generous but plausible split) it will net him perhaps 1% (3% nonvoting, 2% McCain, 1% Obama; a net 1% for McCain).  And this is likely to be cancelled out by a superior Obama get-out-the-vote ground game (which is NOT measured in the polls).  The most likely net impact of these two factors is therefore approximately zero.

There has been no significant movement in national opinion poll averages over the last two weeks. Reported changes in national poll numbers over the last two weeks have been something close to random noise. With multiple polls every day, someone who wanted to make a case for poll movement could cite any number of polls to document whatever movement they wanted to show. But I have looked at hundreds of national polls taken over the last two week: overall patterns have been stable.  If there is to be a last-minute surge for McCain, it hasn’t been measured yet.

One Factor that Could Work for McCain

The Young Vote and Likely Voter Models. All pollsters use a variety of algorithms to determine who is likely to vote and who is not. While all are secretive about the precise decision rules they use, the overall composition is no secret: it is a combination of past voting behavior and knowledge of and interest in the current election. 

But there was a massive increase in primary elections this year in the participation of voters who are under 40 years old. This group favors Obama by a wide margin and there is every indication that they will vote in historically unprecedented numbers in the general election tomorrow.  But we cannot be sure that they will.  If their voting behavior returns to historically typical proportions, it will represent a substantial net vote shift to McCain, enough to make him highly competitive.

The reaction of the Gallup Organization to these changes is especially interesting. They have over 60 years of cumulative experience refining a likely voter model based on past voting behavior and interest in the election. Applying this model has resulted in figures far more favorable to McCain than most other pollsters. But even Gallup is skeptical that their model will work this year.  So, for the first time ever, they are hedging their bets and producing an “expanded” model with an associated estimate in addition to their “traditional” estimate. These estimates drop past voting behavior from the model which has the effect of including more young voters.  The result is closer to the assumptions of most other pollsters. So which will be more accurate this year: their traditional model which has worked well for decades, but which assumes that young people will NOT vote in historically large proportions? Or the expanded model which is based on observations from this year’s primaries and which is similar to the assumptions used by most other pollsters?  The electoral result is likely to depend on which model is more accurate.

McCain’s Best Hope

McCain has improved his position in several of the 13 tossup or Lean Obama states in the last several days. But he has to nearly run the table of all of these states in order to prevail.  If the vote in each state were entirely independent of the vote in every other state, the statistical probability of that happening would be near zero.  But these are not independent events. If something were to cause a last-minute surge for McCain it could well have across-the-board impacts and boost his chances in many states.  But it would have to be a very large surge –approaching 5% --or the reduction in the under-40 vote to historical levels. And he would have to sweep Florida, Georgia, Missouri, Indiana, Ohio AND North Carolina.  All of them. And either  Pennsylvania or Virginia. And some of the remaining tossup states as well.  While it is possible, it seems unlikely that such a significant change would occur this late or that he could win all of these swing states.

 Who Will Win?

I deal in a world of probabilities, not certainties. So I can’t tell you who will win.  But I can provide a “best guess” of probabilities based on what is known.  This is based on a review of hundreds of polls (both national and in every state), trends, and interpretive observations such as those noted above.  My “best guess” probabilities based on all of these factors:

• Obama Electoral Near Landslide (around 350 electoral votes) – 50% probability
• Obama Win by a More Modest Margin – 40% Probability
• McCain Victory – 10%

 Want to know who wins before the networks will tell you?

The networks, for a variety of reasons, are likely to know the results hours before they are willing to broadcast what they know.   I will tell you how to figure out what they know but will not tell you in my Election Night Viewers Guide that I will send out late Monday or early Tuesday.  If you can’t wait for the official network proclamation, look for this guide, and print it out for use on election night. If you do, you will likely know the winner before the network will be willing to tell you who it is.

Michael J. O'Neil, PhD

• Main
• Archives
• Albums
• Links
• Admin

• Census Pick Illustrates Broader Obama Strategy
• Diversity, Perspective, and the Supreme Court
• McMahon Group Sunday May 10, 2009
• Obama's 100 Day Report Card
• Public Furor over AIG Bonuses Misses Bigger Picture
• Got 30 Seconds and Want a Chuckle?
• Jan Brewer and the Arizona Budget Crisis
• Loving and the American Experience
• The Arizona Budgetary Crisis II
• The Arizona Budgetary Crisis in Perspective: What does the public want?

• General [30]

• RSS 0.90
• RSS 1.0
• RSS 2.0
• Atom