March 26, 2012

More Senate Math.......The Overall Outlook for 2012

by Charles Schott

My recent article, Senate Math for 2012....What to expect when your electing!, sets out the context for the upcoming Senate elections.

The article notes that the 2012 Senate races have come into better focus, with the challengers and open seats now mostly identified. It also notes how the basic picture remains very similar to what it was four months ago:

"(I)t appears we are looking at a net shift of between 2 and 8 seats (3 being a Senate tie). My prediction at this point would be for a shift of 3-5 if Obama wins a close race for re-election (which would mean either a tie or R control of the Senate), 4-8 if the R candidate (Romney,...etc.) wins."

The article notes that President Obama recently has experienced several good weeks which included both the unexpected retirement announcement of Republican Sen. Olympia Snowe of Maine and an equally unexpected "coming-out-of-retirement" announcement by former Nebraska Senator Bob Kerry. Each of these announcements creates a competitive Senate races (benefiting the D's) that previously didn't exist.

The article also notes that the economy is now putting up numbers that indicate a tentative return to positive or neutral territory (the rising price of gas and new inflation concerns being notable exceptions).

As a result, D Senate leader Harry Reid recently expressed optimism that his party will keep its majority in the Senate.

The article also expresses the view that if the elections were held today, the R's would likely pick up 2-3 seats, which would not be enough for them to take control of the Senate if there is a second Obama term.

That said, the basic terrain remains unchanged; there is a very real potential for the R's to gain substantially more than three seats and win back the majority they lost in 2006.

There are two helpful ways to analyze and project potential outcomes for the coming Senate elections: (I) at the macro level, using the history of previous Senate elections (particularly those held in presidential election years) to identify a range of likely outcomes consistent with past performance and (II) at the micro level of the individual 2012 races, projecting a range of possible individual outcomes and then looking at them as a portfolio.

I. The Macro Level: Historical experience with Senate races in other Presidential election years.

Looking at the results of prior Senate elections held in presidential election years (see it can be determined, among other things, that:

  • From 1900 to the present, the average shift in Senate seats in connection with Presidential elections has been a win of approximately 3 seats in favor of the party winning the White House.
  • This average three seat swing illustrates how "presidential coattails," as a general matter, have carried along candidates of the winning President's party.
  • While the party winning the presidential election has picked up Senate seats in 18 of 28 such prior presidential elections, there has been a pick-up of 9 or more four times (i.e., 1920, 1932, 1948 and 1980).
  • Incumbent Presidents have been on the ballot in 18 out of 28 races since 1900.
  • Thirteen of these eighteen elections involved a sitting President who previously had been elected while in five the Vice President succeeded to the office and was running for the first time as an incumbent.
  • Only ten times since 1900 has there been an open race without an incumbent President on the ballot.
  • The range of Senate outcomes for all Presidential elections from 1900-2008 is between a gain of 12 and a loss of 4 for the party winning the White House.
  • Similarly, the range of Senate outcomes when Presidential incumbents have been on the ballot is between a gain of 12 and a loss of 3 for the party winning the White House.
  • More importantly, nine out of thirteen incumbent Presidents previously elected and running for re-election have won a second term.

However, the significant numbers are for the range of outcomes not for the winning party, but for the incumbent's party depending on if the incumbent wins or loses:
  • When an incumbent President wins a full term, the range of outcomes is between a gain of 9 and a loss of 3 Senate seats for the winning incumbents party.
  • The average for the winning incumbent's party, however, has been a pick-up of zero Senate seats.
  • When there is no incumbent running, the party of the candidate who wins the White House has had a range of between a pick-up of 10 and a loss of 4 seats in the Senate; The average pick-up has been 3 seats.
  • Since 1900, each of the four times an incumbent President succeeded to the Presidency as Vice President through the death of his predecessor, he won election on his own (i.e., 1900, 1924, 1948, 1964) and brought with him a gain that ranged between between 1 and 9 Senate seats. The average was 4 seats. On none of these occasions did the winning incumbent President lose Senate seats.
  • Gerald Ford (1976) constitutes the one example of a President during this period who succeeded to the Presidency as Vice President through his predecessor's resignation (and de facto impeachment) and the outcome was both a loss by the incumbent President and a loss of one seat by the losing incumbent President's party.
  • Even more interesting is that when an incumbent President wins re-election (i.e., he is not facing the electorate for the first time), the range of outcomes has been smaller, between a gain of 7 (1936) and a loss of 3 (1940); The average, however, is still no change in the Senate seats for the winning incumbent's party, which is surprisingly small.
  • On three occasions when an incumbent President lost re-election, however, the loses to the President's party in the Senate have been sizable for the President's party, i.e. between 5 (1912) and 12 (1932 and 1980). On one occasion the incumbent's party lost only one seat (1992). The average is a loss of 7.5 senate seats. As the earlier article notes, when the economy gets bad enough that the voters want a complete change, the result is typically an electoral tsunami (something Presidents Taft, Hoover and Carter learned the hard way). See

The article notes that from a historical perspective, President Obama, who is running for re-election, falls into one of the last two above bulleted categories. This implies an expected Senate outcome that would range between break-even (i.e., no change if he wins re-election) and a significant loss of five or more seats (if he loses). Interestingly, this is fully consistent with the state-by-state Senate election analysis below.

II. The Micro Level: Looking at the 2012 Senate Races.

Looking at the individual 2012 races, projecting possible outcomes in the individual races, several important things stand out:

A. Underlying the R's opportunity in the Senate is the number of seats the D's have to defend.

"One third of the Senate is up for election every two years and this cycle, of the 33 seats up for election, 23 (including two independents) are being defended by the D's. Additionally, the particular states that have open seat Senate elections in 2012 tend to be more Red on the map than Blue (e.g., MT, SD, WV, NE). All of these listed states went for McCain in 2008 and for President Bush in 2000 or 2004 (both times)."

This is the Senate class that when it last was up for election in 2006 delivered a pickup of 6 seats to the D's along with control of the Senate. See,_2006.

B. The second factor is the 10 seats the R's are defending:

Of seven R incumbents running for re-election, five are races that do not appear to be competitive. These include Richard Lugar (IN); Roger Wicker (MS); Bob Corker (TN); Orrin Hatch (UT); and John Barrasso (WY). Sens. Lugar and Hatch have been subject to speculation resulting from primary challenges from Tea Party backed candidates, so the prospect for additional open seats, although unlikely, still exists. See and

There are also two R open Senate seats which appear likely to remain Republican. These are the seats being vacated by John Kyl (AZ) and Kay Bailey Hutchison (TX).

This leaves three competitive R Senate contests. The two R incumbents considered vulnerable are Scott Brown (MA) and Dean Heller (NV). The open seat is the one currently held by Olympia Snowe (ME).

These three R seats are the likely pool from which the D's will net against any loses from the 23 seats they are defending.

Of the above states, Nevada and Maine now have R Governors (although in Maine this is because of a three way race in 2010). See,_2008 and All three states were easily carried by Sen. Barack Obama in 2008. President George W. Bush did carry Nevada in 2000 and 2004 in close races, while the Democratic nominees overwhelmingly carried Maine and Massachusetts. See,_2000/ and,_2004.

In short, on a presidential level, this group R defended senate seats contains no red states. It is two (deep) blue and one purple.

The R's could win or lose any or all of these three seats. In an election held today they would likely hold one to two. Any that the Republican's lose they need to win an equal number from the D's in the races below just to break even.

C. The individual races for the seats being defended by the D's is the third factor to consider.

Of the 16 incumbent D Senators seeking re-election, 10 currently are considered likely to return.
They are Diane Feinstein (CA); Tom Carper (DE); Ben Cardin (MD); Amy Klobuchar (MN); Bob Menendez (NJ); Kristen Gilbrand (NY); Bob Casey (PA); Sheldon Whitehouse (RI); Bernie Sanders (VT) - Independent Socialist; and Maria Cantwell (WA).

This leaves 13 competitive D Senate contests, seven of which are open and six involving incumbents.

There are seven open Senate seats being defended by the D's, all of which appear competitive. The seven D Senators who have announced they are leaving are: Joe Lieberman - Independent (CT); Daniel Akaka (HI); Ben Nelson (NE); Jeff Bingaman (NM); Kent Conrad (ND); Jim Webb (VA); and Herb Kohl (WI).

Of the above seven states, five (NE, NM, ND, VA and WI) have R Governors. See While five of these states (CT, HI, NM, VA, WI) were carried by Senator Barack Obama in 2008 (see,_2008), only three of them (NE, ND and VA) were carried by President George W. Bush in 2000 and 2004. See,_2000 and,_2004) In short, this group consists of two red states (NE, ND), two blue (CT, HI) and three purple (NM, VA, WI).

Looked at as a group, the R's have a chance in each of these seven races. If the election were held today, the R's would appear poised to win between two (the two red states) and five (the three purple ones at this point could go either way).

There are an additional six competitive races involving D incumbents running for re-election. The incumbent D's currently considered potentially vulnerable are: Ben Nelson (FL); Debbie Stabinow (MI); Claire McCaskill (MO); John Tester (MT); Sherod Brown (OH); and Joe Manchin (WV). The R candidates will for the most part be determined in primaries later this spring and summer, but in some cases the likely R nominees can be seen.

Currently, of the above six states, three (FL,MI and OH) have R Governors and three (MO, MT and WV) have D governors. See Interestingly, all three states with D Governors (MO, MT and WV) were carried by John McCain in 2008 and George Bush in 2000 and 2004. Bush also carried Florida and Ohio in close races both times. In 2012, Florida, Ohio and Missouri are considered swing states and Michigan (a traditionally blue state) is likely to be contested as well. In short: two red states, only one blue and three purple.

The analysis of these six races involving D incumbent Senators running for re-election indicates that if the election were held today, three (MI, OH, WV) currently would be likely remain in the D column, two would be regarded as possible (MO and FL) but too close to call and only one (MT) appearing as likely to move to the R column

Again, while having the theoretical potential to win any of these races, if the election were held today the R's would likely win one to three.


Putting all of the above together (i.e. all the potentially competitive 2012 Senate races), the analysis shows that if the election were held today, there would be a net change in the Senate of between one and seven seats in favor of the R's.

This is calculated by combining the projected current likely outcomes for vulnerable R seats (R's lose 1-2) with the projected results in the open D (D's lose 2-5) and competitive races involving D incumbents (D's lose 1-3).

Even though splitting the difference (i.e., averaging) would result in an R gain of 4 seats, based on the specific states, a more reasonable current projection would be for an overall gain by the R's of 2-3 seats.


The outlook for these Senate races (and also for the Presidential contest) can be expected to ebb and flow in the seven months ahead before the election is finally held on November 6. Additionally, the results in the presidential election will have an impact on the individual Senate races.

The key is appreciating that the level of change in the Congress (both Senate and House) is largely co-variable with the results in the Presidential race. It's not a question of coattails so much as an expression of the polarization in the country and the implications of turnout.

For these reasons, it is worth looking at the range of outcomes that could occur given different presidential election scenarios. This approach makes it possible to look at the set of competitive races as a portfolio where the risks associated with any given outcome are balanced against those in others. This can produce a projection with lower overall volatility (i.e., a greater likelihood of relative stability) and one that is more likely to be closer the eventual result.

A. President Obama Wins Re-election Convincingly.

As stated in my earlier article, the importance of the 1948 Presidential election to President Obama and his team as a model for the 2012 campaign has been made clear. Harry Truman's "come from behind" surprise upset victory over overly confident R's in 1948 is what they would very much like to re-create. See

On the 2012 Senate election side, however, even if President Obama were tp pull off a Truman-like victory for himself, he will not be able to pick up 9 seats the way Truman did. This is because the results of the prior elections for the Senate classes of 1948 and 2012 were quite different. In 1942, six years prior to 1948, the R's had picked up 9 seats, which is the same number the D's won back in 1948 (see,_1942). In 2006, the R's lost 6 seats and as a consequence there are only 10 R seats (out of 33 races overall) being defended in 2012. See,_2006.

In this "Obama Wins Convincingly" scenario, the R's could lose in all three R defended seats (ME, MA and NV) and win only the projected minimum of two open D defended seats (NE and ND) as well as one D incumbent (MT). This favorable result for President Obama would likely result in no change in the current Senate, which would remain split 53 D's' and 47 R's.

B. President Obama Convincingly Loses Re-election.

While recent economic news has been favorable to the White House, the basic economic and financial crisis has not gone away. See As a consequence, a decisive rejection of the Obama Administration and its policies remains one of the possible outcomes of the 2012 election.

As also noted in the earlier article, the presidential elections of 1932 and 1980 resulted in big wins for the challengers (FDR and Ronald Reagan) over previously elected incumbents. These victories carried along a surprising number of new Senators with them.

These two elections (and also 1912 when Woodrow Wilson defeated both President Taft and former President Teddy Roosevelt) mark the watershed Presidential elections of the twentieth century when the American people rejected incumbent Presidents, their parties and the status quo. When the economy gets bad enough that the voters want a complete change and the result is an electoral tsunami. Presidents Taft, Hoover and Carter learned this the hard way.

In this "Obama Loses Convincingly" scenario, the R's would likely lose only one R defended seats (ME), while the D's could lose as many as six open D defended seats (NE, ND, NM, VA and WI and either CT or HI) and conceivably lose an additional three D incumbents (MT, FL, MO). This would result in an R pick up of eight seats in the Senate, and an R majority of 54 R's' to 46 D's.

These two scenarios appear to represent the range of possible outcomes for 2012 within the current outlook. Note that both of the above scenarios are consistent with historical precedent set out above; prior Presidential re-elections point to a likely gain of zero if the President wins and a loss of five or more if he loses.

To get the most likely expected outcome for the shift in the Senate would require running a series of simulations after assigning probabilities for different possible expected outcomes (and co-variances) for both the presidential and individual senate races (see

Absent that, the above "if the election were held today" projection of an R gain of 2-3 seats appears reasonable.

A word of caution is appropriate. Just as there is no such thing as a free lunch in economics, there are no sure things in politics (where "a week can be an eternity"). Things that look reasonably certain can change suddenly, either nationally or locally. Scott Brown's election in Massachusetts in 2010 to fill the late Ted Kennedy's seat is a good example of an unexpected development. Olympia Snowe's unexpected retirement announcement at the end of February is another example.

That said, there typically are constraints and probabilities governing the possible outcomes and it is helpful to focus in on what is likely "within the twenty yard lines."


For the R's, a good night in 2012 will mean a pick-up of four or more seats, which would mean R control of the Senate.

To achieve this, the R's would need to lose one of the three competitive R races and win five of thirteen competitive D seats.

This combined with an average performance in 2014 (when the D's will have to defend the 20 D Senators elected with President Obama in 2008) would likely cement R control of the Senate until at least 2016, and possibly until 2018 when the 2012 class will be up again.

In this scenario, with a new Republican president, the R's have a chance to break the gridlock that has characterized Washington.

The current "if the election were held today" projection of an R pick-up of 2-3 seats falls just short of that.

That said, given the expected ebbs and flows between now and November, the possibility of an R outcome that breaks the gridlock is quite real.

The alternative is the current gridlock in Washington continuing for another four years.

Charles Schott served in the last three Republican administrations and served as Deputy Assistant Secretary of the Treasury in the George W. Bush Administration. He can be reached at




Return to



Site Meter

(c) 2006-2012