Pages

Tuesday, November 2, 2010

11 PM: Updated House Estimate GOP +63/+67 Median Price Estimate +64

11:15 PM model run based on InTrade data suggests seat range of 240 to 245 with a midpoint estimate 241.  More analysis to follow later.

Monday, November 1, 2010

House GOP Plus 62 ... or More; Senate GOP 49+










House Republicans go into election day in their strongest position of the election season according to our model's evaluation of InTrade market data  (data download at 8:00 pm).  Republicans go into election day with a 92.8% probability of capturing the House of Representatives.  Our model suggests that they will emerge with at least 238 seats in the next Congress -- a gain of 61 over their post-2008 election total.  The midpoint estimate in our model would give Republicans a total of 239 seats and the upper end pricing point would yield 240 seats.  Data suggests that Republican chances of exceeding 65 seats gained are approaching 50%.   NOTE:  Updated Estimates will be posted throughout the day, even after the polls close.

On the Senate side, our micro and macro models are both projecting a Republican gain of at least eight seats.   Our micro model projects an upper limit gain of 8 seats for a total of 49.  Our macro model's upper limit pegs the gain at 49.5 seats. 


Seven states have defined the main battleground over the past couple of weeks.  Republicans are ahead in the probabilities in four of the seven (they need 6 for control) and Democrats in three. 

Tomorrow promises a great deal of volatility. The question is where we will see the volatility. Watch for volatility especially after noon Eastern time when the cybershpere will be overloaded with rumors concerning turnout, voter problems, and supposedly leaked exit polls.



We will offer our views of what the exit polls and early returns portend as soon as we get our hands on the data.  Check back around noon tomorrow for our detailed election viewing guide and further prognostication. 

Last evening our post noted that we expected election day to arrive with  the Neither Party contract overtaking the Democrat control contract in InTrade trading.  Our sense of the market trend has proven to be on the money.  We would expect to see continued declines in the Democrat control contract and gains in both the Republican control and Neither Party control contracts.

Throughout the evening we will also post updated probabilities for Republican control of the Senate based upon decided races.

The table below outlines the probabilities of Republicans gaining a specified number of the seven seats or more.  This is one of the metrics that will be updated throughout tomorrow evening.



For tonight, a couple of other charts are also pasted below that are self explanatory.  The House control chart, in particular, shows how the erosion of Democrats' position has deteriorated since passage of the Health care reform legislation.  Also, make sure you check out our updated InTrade play portfolion on our InTrade portfolio page.






Sunday, October 31, 2010

Expect More Than 60 Seat Republican Gains in House









Republicans are poised to gain more than 60 House seats this Tuesday from their post-2008 election total.  Our model currently projects that Republicans will hold 237 or 238 seats in the next Congress.  The probability of a Republican House takeover currently stands at 92.5%.

All of our model projections are based on a statistical analysis from InTrade prediction markets.  The model estimate will be updated not just Monday and Tuesday morning, but will continue to be updated throughout election day and night.  In fact, should races remain unresolved for several days, our models will continue to provide updated information until all the contracts have settled.

While the House cake appears to be baked, or at least we're checking with a tooth pick, the Senate is a different story entirely.

Of the seven races that most analysts agree will decide Senate control, Republicans are well positioned to capture four of those seats.  To gain control of the Senate, Republicans must win six of the seven -- or pull an upset someplace else in the country.

As the chart to the left and table above show, it appears to be a tall order for Republicans.  Republicans would need to win two of West Virginia, Washington, and California.

Our models suggest that Republicans are poised to hold at least 49 seats in the Senate.  This sets up an interesting situation for traders of InTrade's Senate control contracts.  As a reminder, Independents Lieberman and Sanders do not count toward control for Democrats, but Lisa Murkowski would count toward Republican control.  Bottom line -- if Republicans hold 49 seats, Democrats would hold 49 and there would be two independents meaning that the Neither party control contract would prevail. 

If the seven races are all that matters (everything else goes as expected), Republicans would have a 5% chance of taking control of the Senate.  If Pennsylvania is called for the Republicans, their chances improve marginally to 6%.  A West Virginia win would improve the chances to 18%.  On the other hand a loss in West Virginia would drop their probability of Senate control to under 2%.

As each Senate race is decided on Tuesday, updated probabilities will be posted here on the blog.

Over the last few days we have seen the beginning of a convergence between the Democrat control contract and the Neither party control contract.  Republicans have also seen a slight uptick, but remain at a sub-20% probability at this writing. 

The trend in our model suggests that it is a distinct possibility that we will see the Neither party control (under InTrade contract rules) favored by election morning.   As mentioned earlier, Republicans appear poised to hold at least 49 Senate seats in the next Congress.  Both our macro and micro models suggest 49 as the number to expect.

However, there is a real possibility that this estimate will change over the next 36 hours.

There were several polls released over the last 48 hours in a number of the Senate races.  While our models only look at the InTrade data, traders do look at the polls as part of their trading strategies.  Several polls suggest that Washington state will be a real barnburner that may take weeks to resolve.  If this turns out to be the case, it will be interesting to observe trader reaction to each new vote total as it is released.

One poll in Sunday's set of polls did catch our attention --  Public Policy Polling's (PPP) West Virginia offering which showed Joe Manchin with a 5 point advantage.  Rasmussen's recent poll  showed Manchin +3.  PPP's offering is quite reasonable and Election Markets isn't suggesting that it is an outlier.  What is striking is the response to PPP's question about 2008 vote for President.  Senator John McCain carried West Virginia by a margin of 56% to 43% (+13).   PPP's respondents said they voted for McCain 52-40%.  The 12 point margin is roughly the same as the actual result.  What's surprising is that their likely voter model seems to be implying that the West Virginia electorate's makeup will be roughly the same as 2008.

Exit polls in 2008 pegged the electorate's partisan makeup as 48% Democrat, 34% Republican and 19% Independent.  The PPP poll's likely voter composition is 55% Democrat, 34% Republican and 11% Independent.  If you apply PPP's partisan responses to the 2008 exit poll partisan makeup you get a 49%-49% tie rather than a five point lead for Manchin.  This PPP poll and their prior one from one week earlier used the same Likely Voter partisan makeup.  However, their September 21st poll used a partisan makeup much closer to the 2008 exit polls.  It's LV makeup was 51% Democrat, 37% Republican, and 12% Independent -- if you apply that electoral makeup you get 49% Manchin and 48% Raese result.

PPP is a solid pollster.  Not saying they will be wrong.  But these pages have questioned whether or not the big surprise of this election will be the performance of likely voter models.  We find it difficult to believe that the partisan makeup of the electorate will be more favorable in virtually any state than the electoral makeup in 2008.

This is a key race to watch on election night.  Fortunately, it will be one of the earlier races all of us will get to see data on.  You can download our election night viewing guide starting at about noon (EDT) on Tuesday.  It will provide our view of key races in the House and Senate to watch for clues to the final outcome.