Tag: incumbency

We know money and several other factors have major effects on House races.  But after we account for these major factors, how much advantage does incumbency give a candidate?  A gerrymandered district?  Getting caught in a scandal?

Yesterday I showed some regressions for Republican performance in House races for the years 2002, 2004, and 2006 that take account of incumbent party, fundraising ratio, and district partisan makeup.

Using these, we can tell how well we expect a Republican to do given certain conditions.  However, the regressions are not perfect – the data don’t fall along the lines plotted.  There’s plenty of room for other factors to be involved.  We can use the differences between what we expect and what actually happened – the residuals – to tease out the effects of additional conditions.  Below, a pack of factors, from the most important – money, party, district – to the less important ones – incumbency, gerrymandering, longevity – to the more interesting ones – scandal and failure.

Cross posted at Open Left and Daily Kos

How do various factors affect a House candidate’s percentage of vote in the election?  All the following numbers relate to average effects.  Individual results may vary.

The first four are the variables used to predict the expected performance:

1.   High D/R Fundraising Ratio:  +15 points (about 5 points for every factor of 3 increase)

On average, challengers running for a seat currently held by the opposite party will gain 15 points if they outraise their opponent by a factor of three compared to if they raise only one tenth of their opponent’s money, if all other factors are equal.  Note that the D/R Fundraising Ratio is fairly closely correlated to absolute amount of money raised by the challenger, so we can also say that challengers will greatly improve their performance if they raise a great deal of money, regardless of the incumbent’s fundraising.

2.  Running as incumbent party:  +10 points

Candidates running for a seat currently held by their own party (incumbents or open seat candidates) will gain, on average, 10 points compared to if they were running for a seat currently held by the opposite party, if all other factors (including D/R Fundraising Ratio) are the same.

3.  Running in a more favorable district:  +10 points (1 point for every 3 point change in Bush’s vote in the district)

Democrats running in the most liberal Republican-held districts (Bush vote 40-45%) will gain 10 points compared to Democrats running in the most conservative Republican-held districts (Bush vote 65-75%), if all other factors (including D/R Fundraising Ratio) are the same.  Democrats running in the moderately liberal Democratic-held districts (Bush vote 30-35%) will gain about 10 points compared to those running in the most conservative Democratic-held districts (Bush vote 60-65%).  

4.  Political climate:  +6 points

On average, Democratic challengers did 6 points better against Republican incumbents in 2006 compared to 2002 (4 points better than 2004) when accounting for D/R Fundraising Ratio and district partisan makeup.  Republican challengers did 4 points worse in 2006 compared to 2002.  In other words, Republican money was worth less in 2006 than in 2004 or 2002.  They had to raise more relative to their Democratic opponent to get the same result.

The following comparisons are made by comparing actual performance to calculated performance, accounting for the four factors above: D/R Fundraising Ratio, district composition, incumbent party, and political climate.  The numbers given are average residuals of the regressions.  

5.  Raising more than $2 million as a challenger: +3 points

Remember, this is after accounting for D/R Fundraising Ratio.  If both candidates raise the same amount of money, dollar-for-dollar, then the more money a challenger raises, the better the challenger does.  A challenger who raises more than $2 million (and whose opponent also raises more than $2 million) increases performance by about 3 points compared to one who only raises $100,000 (and whose opponent also raises only $100,000).  In other words, high-spending races with fundraising parity are generally to the advantage of the challenger.  (This leads to the strange corollary that the more an incumbent raises given fundraising parity, the worse the incumbent does!) Let me note again, when we do not control for D/R Fundraising Ratio, a challenger who raises a large amount of money will do far, far better than one who raises little money.

6.  Running as an incumbent:  + 2 points

The inherent incumbent advantage after accounting for money, party, district, and climate is not large.  This doesn’t mean running against an incumbent is just as easy as running for an open seat.   However, the incumbency advantage may reside mainly in the ability to scare off opponents and scare off opponents’ donors and supporters.  If a challenger can manage to raise as much money as an incumbent, then the challenger has almost as good a shot as if the challenger were running for an open seat.  However, 2 points is still an important amount.

7.  Running against a first-termer: +1 point

First term incumbents are not much more inherently vulnerable than other incumbents, if at all.  Even those who are in a seat that switched parties.  This doesn’t mean first-termers are safe, because they are more likely to attract high quality opponents with strong fundraising.  When they do, however, they perform only slightly worse than a long-time incumbent under the same circumstances, on average.

8.  Running against a self-funded candidate:  +1 point

On average, running against a self-funded candidate might give a slight advantage.  However, out of the 18 cases I found over the past three cycles, four showed the self-funded candidate underperforming by a massive 8-10 points.  There may be a risk of completely blowing it by self-funding.

9.  Running against a Republican incumbent in a Republican-gerrymandered district: +0 points

Looking at some states that were recently redistricted by Republicans in a partisan manner – FL, PA, MI, OH, VA, TX – there has been no benefit in performance for the Republicans.  There may have been a slight benefit the first cycle after redistricting, followed by a slight underperformance later.  The gerrymandering may have scared off opponents and their donors, however, which would certainly have been an overall benefit for the Republicans.

The following comparisons are specific to a just a few races, so we run into the problem of the statistics of small numbers, and can’t really say what the average effect is.  Also, in many of these races, the incumbent was tangled in more than one variety of misdeed.

10.  Third party candidates:  0 to -15 points

In 2006 there were 16 House races where third party candidates garnered more than 4.5 percent of the vote.  In 11 of these races the Republican underperformed by 4 or more points; in 6 races (2 in MN) the Democratic candidate underperformed by 4 or more points.

11.  The Abramoff scandals:  -1 to -12 points  

Republicans in districts with links to the Abramoff scandal all underperformed: TX-22 (-1),  FL-24 (-3), CA-4 (-4), AK-AL (-6), CA-11 (-7), and OH-18 (-12).

12.  Alleged domestic abuse:  -5 to -6 points

PA-10 (-6), NY-20 (-5):  Not the good kind of press.  

13.  Threatenting your opponent:  -5 points

WY-AL (-5), where Barbara Cubin told an opponent she’d slap him in the face if he weren’t in a wheelchair.  Cubin wasn’t well liked anyway though.

14.  The Delay scandal:  +5 to -6 points

TX-22 (-1), AZ-1 (-2), NC-8 (-6), PA-6 (+5).  Districts related to the Delay scandal don’t seem to have been affected too much, although the Delay scandal certainly affected the national climate.

15.  The Foley scandal:  +1 to -3 points

IL-14 (-2), IL-19 (+1), FL-16 (0), NY-26 (-3).  Again, no obvious severe penalty for those most closely related to the scandal or Foley’s replacement on the ballot, but the scandal contributed to the national political climate.

Overall, these numbers seem to validate the strategy of supporting strong candidates in every district, against every incumbent.  While it is certainly much more difficult for Democratic challengers to win against an incumbent in a conservative district, it is not impossible.  It appears that with enough money, such races will often be competitive or near competitive in the current political climate.  Another way to put it is that the competitive races in conservative districts in 2006 -WY-AL for example- were not simply flukes or outliers, but rather part of a larger pattern that is likely to be repeated in 2008.  

Thanks to Open Secrets, fundraising data is readily available for all House candidates.  This diary sorts through all the House races from the last three cycles to show the effects of money, incumbency, political climate, and party on the elections.  

While money was clearly the most important factor, the big surprise was that once money was accounted for, running against an incumbent was only a little more difficult that running for an open seat, on average.

The effect of relative fundraising strength (the D/R Fundraising Ratio, Democratic $$ raised divided by Republican $$ raised) for all 2006 races is below:

Click to enlarge.

On the far left, when the Democrat raises very little money compared to the Republican (D/R Fundraising Ratio < 0.1), the Republican always wins.  On the far right (D/R Fundraising Ratio > 10), the opposite case.  Much much more below…  

Cross posted at Daily Kos and Open Left.

Sorting Through the Data

First, notice the scale on the bottom of the graph above (and all of the ones to follow) is not linear (1, 2, 3,….) but rather logarithmic (1, 10, 100…).  This is a way of showing a very large range of values on one plot.  The line on the plot above is a smooth curve through the data.  Notice that the curve flattens out at either end – these are the areas where one opponent basically has diddle squat.  It doesn’t make much difference whether you have diddle squat or diddle squat times ten, you still can’t run a very effective campaign (although intriguingly there appears to be some effect).

But wait – is the shape of this curve influenced by the distribution of the partisan makeup of districts?  There is an imbalance in district distributions – far more are very liberal than are very conservative.  So, let’s look at only districts where 50-55% of voters chose Bush in 2000, a particularly large group.  

Click to enlarge.

It’s still a pretty nice curve.  A little lopsided though – so should it really be just one curve?  Let’s plot the data by party of the incumbent:

Click to enlarge.

Party makes a pretty big difference.  

Finally, that flat part – out where somebody is raising diddle squat – just isn’t very interesting if we want to know about the effect of money in competitive or near-competitive races.  Let’s narrow things down to a range that appears to be relatively linear on these plots:

Click to enlarge.

Now we can put up some regressions, instead of just smooth curves.  And, if we add open seats formerly held by Republicans, we see something interesting:  open seats look pretty similar to seats with incumbents in their behavior.  More on this in a minute.

So we see this:  the more money a challenger raises compared to the incumbent, the better the challenger does.  Not shocking, but we have a chicken-and-egg problem:  Do challengers raise more money against some incumbents simply because the incumbents are more unpopular?  In which case, would it be because the incumbents are more unpopular that they are doing worse, not because the challengers are raising more money?

Are Popular Incumbents Vulnerable to Well-Funded Opponents?

To check this, I tried to find Republican incumbents with decent challenges in 2006, who were nonetheless considered to be popular or relatively safe in their districts, and who did not have a strong challenger in either 2002 or 2004.  I used some advice from a couple House experts, polls, and local news sources to choose the following districts:  DE-AL, IA-2, MN-1, MI-8, OH-1, PA-4, VA-10, MT-AL, CO-6, CA-2, KY-2, IN-3.  I could have chosen poorly on some of them but hopefully not all of them.  Here’s a plot of the Republican performance in all these districts over three cycles, adjusted for the national mood (more on that later):

Click to enlarge.

If all the popular incumbents were truly Teflon-coated, then we would expect all the points for all years to fall on a flat line.  Instead, they Republican performance decreases with increasing D/R Fundraising Ratio, just like all the other seats.  Plotted individually, this is also the case for 10 out of the 12 districts (3, randomly chosen, shown here).  In other words, a popular incumbent can be defeated with enough money, just like anybody else.  It is getting the money and the candidate that is the hard part.  

The 2006 Election

So then, let’s look at the 2006 data for Republicans in the competitive range I defined above, divided into categories based on Bush’s support in the district in 2000:

Click to enlarge.

As you can see, in each case there is a decent trend, but it’s not very tight.  In other words, after accounting for the D/R Fundraising Ratio and the partisan makeup of the district, there’s still plenty of room for other factors, including, quite simply, the quality of the candidates.  

It’s also pretty clear that the open seats (which were not included in the regressions) aren’t too different from the seats with incumbents after accounting for the effect of money.  We also see, of course, that the open seat contests are far more likely to have fundraising parity (D/R Fundraising Ratio = 1) or better, which explains (mostly) why they are far more likely to change parties.  In other words, incumbency matters for getting money and chasing off opponents and opponents’ supporters, but if a well-funded opponent shows up anyway, that opponent has nearly as a good a chance (on average) as if they were running for an open seat.

Another interesting thing is that the plots don’t look too different from each other.  

The Last Three Cycles

Let’s look at all the regressions for Republican incumbents, and add in the Democrats, for 2002, 2004, and 2006:

Click to enlarge.

It’s hard to tell the lines apart in the graphs, but that is my point.  There is surprisingly little difference between the regressions!  Some of the lines are a little funny looking, but they are the ones based on only a small number of points (especially the ones for the most liberal districts).  Republicans in very conservative districts perform better than those in moderate districts, given the same D/R Fundraising Ratio, but not 30 points better.

We can also see a difference between the years.  The regressions generally shifted down in 2006 compared to 2002 or 2004.  This means that the Democrats’ money was more effective in 2006, and the Republicans’ money was less effective.  This is why predicting electoral success based on fundraising alone, extrapolated from 2002 and 2004 results, failed.

And, of course, the party of the current occupant makes a difference.  If both candidates in a race raise the same amount of money, a Republican incumbent, or a Republican contestant in a Republican-held open seat, will (on average) score a good deal higher than a Republican challenger in a district with a similar partisan makeup.

Conclusions

Tomorrow, I will continue looking at the House races by exploring 15 factors and how they affected candidate performance between 2002 and 2006.  In the meantime, we can say the following:



1.  Money matters a great deal.  

2.  Safe incumbents are not really safe if a well-funded challenger runs against them.

3.  Voters have a habit of voting for the party of the current representative, whether the incumbent is running or not.

4.  The partisan make-up of the district matters, but not as much as one might have expected.

5.  The national political mood matters, and right now, it is still surging against Republicans.