The New York Times published a fascinating article last week about the changing politics of physicians. The article reports that as more physicians abandon practices for salaried positions their politics have trended to the left. This trend is also present in the contribution patterns of health care professionals.
The figure below tracks the changes across time in the ideological giving patterns of four groups of health care professionals: surgeons, nurses, mental health care professionals (restricted to psychiatrists and clinical psychologists), and all other physicians. The trends track the mean ideological position (ideal point) of donors within each group. For each election cycle, I take the subset of donors from each group that gave at least once during that cycle and calculate their ideal points based on all donations made during that cycle and cycles prior. This permits the ideal point estimates to update with time.
Here are some helpful summary statistics for interpreting the scale. The mean ideal point for the entire sample of 2.5 million individual contributors (which includes health care professionals) is -0.16. The mean ideal point is -0.66 for Democratic candidates and 0.82 for Republican candidates.
Surgeons are by far the most conservative. They are also the group that was, on average, more conservative in 2010 than in 1990. The other three groups trend to the left in varying degrees.
I included nurses to emphasize that the industry has been changing as a whole. I also included psychiatrists and clinical psychologists to emphasize that differences between groups is probably better understood as a function of self-selection into a given profession than anything else. Whatever differences in pay and work conditions exist between mental health professionals and other physicians, they would not be sufficient to explain the differences in ideology. That mental health professionals are significantly to the left of nurses reinforces this point.
Although revealing, the trend line for physicians leaves out much of the story. The mean ideological position is relatively stable, but there is plenty of action in the distribution. The 1994 and 2010 midterm election cycles represent the closest thing to electoral deja vu that we can expect to see in our lifetimes. This makes these cycles useful points for comparison. The figures below show the ideological distributions for surgeons and physicians in 1994 and 2010. The distribution for surgeons fills out a little on the left over the years but remains unimodal. In contrast, the distribution for physicians changes quite a bit. Ideologically speaking, physicians have become more evenly divided and more polarized.
These figures can help address one of the paragraphs from the article that caught my attention:
“Dr. Cecil B. Wilson, the president of the A.M.A., said that changes in doctors’ practice-ownership status do not necessarily lead to changes in their politics. And some leaders of state medical associations predicted that the changes would be fleeting.”
I strongly suspect that Dr. Wilson is correct in his assessment. The link between doctors abandoning private practices for salaried jobs and their changing politics is probably overblown. I attempted to get at this question by dividing the sample into physicians who report being self-employed and physicians who report having an employer. I found only a slight difference between the groups. On the other hand, I sincerely doubt that the changes will be fleeting. The changes appear to be responding to generational shifts—in particular, the influx of women doctors.
I was curious as to whether the rightward jump observed during the 2010 midterms resulted from 1) donors that had previously given to Democrats shifting their dollars to Republicans or 2) increased giving by Republican donors relative to Democratic donors. In other words, was it a case of changing minds or changing wallets? For each year, I categorize donors into one of six categories based on their giving patterns in previous cycles.
- Strong Democrat – greater than 95% of donation dollars went to Democrats in prior elections
- Lean Democrat – between 60% and 95% of dollars went to Democrats in prior elections
- Toss Up – between 40% and 60% of dollars went to Democrats in prior elections
- Lean Republican – between 5% and 40% of dollars went to Democrats in prior elections
- Strong Republican – less than 5% of dollars went to Democrats in prior elections
- New Money – first time donors/had not given in prior elections
Each bar displays aggregate amounts donated to each party. For example, the bar labeled “Strong Dem” in the 2010 panel shows that donors who in previous cycles gave 95% or more of their dollars to Democrats, gave $41 Million to Democrats and $1.5 million to Republicans.
Partisan defections by Democratic donors had little to do with the 2010 shift in favor of Republicans. (In fact, the defection rate was greater for Republicans donors.) Increased giving by Republican donors and slightly reduced giving by Democratic donors account for the lion’s share of the swing. Republicans also received a larger proportion of dollars from first time donors than they had in the past. Republicans won 62 percent of the dollars from first donors in 2010, up from 44 percent in 2008 and 57 percent in 2006.
Lastly, it is worth noting that the overall rightward shift during the 2010 midterms is no larger than what we saw in other industries. Hence, I caution against concluding that the shift was in direct response to the Affordable Health Care Act.
Now that the dust from the 2010 Elections has cleared, it is time to start looking at how the 112th Congress will compare to the 111th. The figure below tracks the ideological positions of the median House member, the mean Democrat, and the mean Republican since 1988. The replacement effect of the 2010 Midterm Elections is unlike anything in recent memory. The shift in the House median is two and a half times what was observed after the 1994 Election, wiping out the effect of Democratic gains in the previous two elections and then some. The 111th was the most liberal Congress in the past three decades; the 112th will be the most conservative.
The 2010 Elections had a profound effect on congressional polarization. Not only will the 112th House be the most polarized on record; 2010 will surpass 1994 as the most polarizing election cycle.
The figure below helps place the historic jump in polarization in perspective. It displays the ideological distributions of House members continuing on from the 111th Congress, alongside the distributions of the entering class of Republicans and the sitting Democrats they defeated. Two features stand out. The first is that the mean entering Republican (1.09) is substantially more conservative than the mean continuing Republican (0.82). In total, 77 percent of freshmen Republicans in the 112th Congress will locate to the right of the party median from the 111th. In other words, nearly 8 in 10 incoming House Republicans would have been on the right wing of the party in the 111th Congress.
The second standout feature is that, contrary to many media accounts, the Republican wave was an equal opportunity un-employer for Democrats. Democrats who lost their reelection bids were slightly more moderate than those who retained their seats—0.61 compared to 0.64—but the difference is statistically insignificant. The attrition rate was greater among centrist Democrats, but there were so few Democrats remaining in the political center that the losses had to come from elsewhere.
The polarization resulting from the 2010 Midterms is fundamentally different and more worrisome than what had preceded it. By historical standards, the post-war era stands out as a period of relatively low partisan polarization. This is largely attributable to the coalition between Northern and Southern Democrats. The increase in polarization during the 103rd through 105th Congress corresponds to the tail end of the Southern partisan realignment, a period during which southern districts that had traditionally elected moderate Democrats (a.k.a. Dixiecrats) began electing conservative Republicans. As the Southern Democrats gradually disappeared throughout the 1980s and 1990s, the parties became more clearly defined, thus returning congressional polarization to the historical norm.
The hollowing out of the political center explained the momentous rise in polarization during the Southern realignment. Now that only a handful of moderates remain in the House, polarization can no longer be portrayed as a story of vanishing moderates. It appears the rise of the extremists has stepped up as the driving force behind congressional polarization.
As promised, here are my predictions for the 2010 Midterm Elections. The model parts with poll-based forecasting models in predicting that the Democrats will maintain control of the House. It predicts that Democrats will win between 217 and 238 seats, which translates into a loss of 19 to 40 seats. (The complete set of House predictions is available for downloaded as a .csv or a .xls.)
If Congress operated more like a Westminster parliamentary system, fixating on which party will win a majority of seats would be more sensible. In such a setting, after assuming power the majority party (or coalition) is free to enact legislation with as little input from the minority as it pleases. This is not the case in Congress. The past two years have been a constant reminder to Democrats that even large electoral majorities do not grant similar levels of legislative control. Current theories of congressional behavior tell us that the position of the median member of Congress can be as important to policy outcomes as which party is in the majority. An advantage of my forecasting model is that it can predict ideological quantities of interest other than seat shares. For example, I can predict the position of the median member in the next Congress and the extent to which partisan polarization will increase or decrease.
The model projects that the position of the median House member in the 112th Congress will be -0.05 with a 95 percent CI between -0.13 and .13. This represents a sizable shift to the right from the median legislator in the 111th Congress, who was located at -0.24. Yet, this will only bring the median back to where it was during the 110th Congress. To place this in perspective, the median House member in the 111th Congress was in the region of Joe Baca (D-CA) and James Oberstar (D-MN). The model predictions place the median member for the 112th Congress in the region of Arthur Davis (D-AL) and James Marshall (D-GA) but could be as far to the right as John McHugh (R-NY) or former Senator Lincoln Chaffee (R-RI), which is still very moderate. According to the model, even in the best-case scenarios, the House median will be much more moderate than what Republicans experienced during the 104th-109th Congress.
The figure below displays the trend lines for the median House member and the means for each party since 1990. Regardless of which party claims a majority after the election, the model projects an increase in partisan polarization. The mean Republican will experience its largest shift to the right ever recorded, while the mean Democrat also will move further to the left as Republican challengers pick off moderate Democratic incumbents. The general rule of thumb for this election is: the larger Republican gains, the greater the increase in polarization.
I report the model predictions faithfully here, but I remain somewhat skeptical of the model predictions for two reasons. The first is that the realm of campaign finance has undergone changes since the previous election cycle. Not accounting for independent expenditures by outside groups might have biased the model in favor of Democrats. On the other hand, the BCRA arguably represented a much larger shock to the campaign finance system than Citizens United, yet the model predictions for the 2004 Elections were right on target. Moreover, the model does not include any variables that relate to campaign expenditures; it only conditions on fundraising patterns, which remain largely unaffected by Citizens United. The second reason is that the model predicted Democrats would win about 10 seats fewer than they actually did in the 2006 Midterm Elections. It is difficult to determine whether this reflects the Mark Foley October surprise or a failure by the model to account for partisan momentum.
I suspect the model predictions are too generous to Democrats by about 8 to 12 seats. Even with the downward adjustment, I still predict the Democrats will retain their majority, but just barely. This is in line with Sandy Gordon’s forecasting model based on calibrated expert raters. Along with Sandy’s forecasts and the recent polls showing that oversampling of landlines can bias polls in favor of Republicans, my model provides additional evidence that poll-based forecasting models are overstating Republican gains. Fortunately, we won’t have to wait long to find out.
As election day nears, I thought it might be an interesting exercise to see how accurately I could forecast election outcomes using only information derived from campaign finance records. Campaign finance records represent a rich data source that speaks to many areas of U.S. politics. Elections are no exception. I’m not the first to incorporate information on fundraising into a forecasting model. However, to the best of my knowledge, I am the first to attempt to forecast election outcomes based solely on FEC records. Despite the handicap of excluding all information from polls, InTrade, expert raters, and other data sources used to forecast elections, the model’s predictions are remarkably accurate. In fact, the out-of-sample predictions for House races outperform the polls.
Contribution records contain far more useful information than what can be expressed by fundraising tallies. For instance, they provide a way to estimate a reliable set of candidate positions via CFscores. The CFscores update almost in real time as FEC records are released during the course of an election cycle. In other words, we don’t have to wait until the election is over to get ideological measures for candidates. The CFscores enable my forecasting model to account for factors that other models ignore, such as adjusting for whether the ideological extremity of Tea Party candidates will hinder Republican electoral prospects this November. In addition, I can forecast ideological quantities of interest such as how the location of the median member of Congress will change after the election. This is perhaps less useful in terms of the horse-race but is probably a better overall measure for the type of policy we should expect from the next Congress.
Also informative are the patterns of individual donors across elections. I’ve been working on assigning unique contributor IDs that link contribution records from the same donor across election cycles and across state and federal elections. This may not seem like a big deal, but the ability to track the behavior of individual donors across elections cycles unlocks a wealth of information that had previously been trapped inside the dataset. For instance, linking records across years makes it possible to calculate the proportion of a candidate’s funds that came from first-time donors as opposed to veteran contributors. At the level of campaigns, this can convey information about a candidate’s success in activating supporters. At the national level, the median CFscore across all first time donors serves as a good proxy for the enthusiasm gap.
Much of the model’s predictive power is owed to an idea I borrowed from Sandy Gordon. Sandy has an interesting paper in which he uses past performance of expert election raters to calibrate their predictions in future elections. The paper gave me the idea of treating the hundreds of thousands of donors who had given in previous election cycles as de facto expert raters by looking at the percentage of funds given to winning candidates in previous election cycles. The idea behind this is simple. Some contributors give a greater proportion of their money to candidates that go on to win, while others spend the majority of their money on losing candidates. All else equal, the more money a candidate raises from the type of donors who give to winners, the more likely he is to win. (For additional details on what goes into the model, I include at the bottom of this post a description of the other model predictors, as well as links to download the R script and dataset.)
Tom Holbrook over at Politics by the Numbers nicely overviews the accuracy of election polls. He shows that during the 2006 and 2008 election cycles approximately 85 percent of House candidates who led in the polls 45 days before the election went on to win. As a comparison, the out-of-sample predictions from my forecasting model correctly identify the winner in over 94 percent of combined 2002-2008 House elections.
It is worth noting that my sample includes a number of less competitive House races that lack polling data and hence are excluded from the poll based predictions. The larger sample accounts for some of the increase the prediction rate but not all of it.
The model’s seat share predictions are also close to the mark. The table below shows the number of seats won by Democrats that the model predicts above/below the observed outcome (e.g. a value of 4 indicates the model predicts that Democrats would win four more seats than they actually did; a value of -4 indicates the model predicts Democrats would win four fewer seats than they actually did). I ran 1000 bootstrapped simulations for each election cycle to get uncertainty estimates. The first column reports the median value from bootstrapped runs and the other two columns display the upper and lower 95 percent confidence bounds.
CI Lower Bound
CI Upper Bound
The predictions are very close to the actual outcome in all but one election cycle. In 2006 the model under predicts Democratic gains by a considerable margin. This might be a result of the Mark Foley factor but I would need more evidence to support that claim. It is just as likely that it reflects a failure of the model to adjust for the effects of partisan momentum in landslide years. (As a note, I noticed that the confidence bounds only contain the observed value in two of the four elections. This suggests that I should probably up the amount of uncertainty in the bootstrapping scheme above the software’s default.)
A major advantage of using campaign finance data to forecast elections is that it costs next to nothing. There is no need to commission polls or pay expert raters. One needs only to collect freely available data and fit a model. That being said, campaign finance based forecasting could have the greatest impact in state level elections where polling data is sparse but fundraising abounds. Although I’m not convinced that they would be, it might be interesting to seeing if the model predictions are as accurate for state elections as they are for federal elections. The catch is that not every state is quite up to speed with the FEC in releasing contribution records to the public in a timely manner, but this problem is fast solving itself as the disclosure process becomes increasingly digitized.
As much as I would like to have predictions ready for the upcoming election, I haven’t had the time to fully organize the dataset. Those forecasts will have to wait until this weekend. In the meantime, I’ve made the dataset and R script available for download for anyone who might be interested.
Candidate Positioning: I use candidate CFscores to measure ideology.
Picking Winners: For each contributor, I calculate the percentage of funds given to winning candidates in previous election cycles. Then for each candidate, I calculate the mean value of his contributors weighted by the dollar amounts received.
First-Time Donors: I calculate the percentage of a campaign’s donors that gave the first time during that election cycle. This is intended as a proxy for the enthusiasm gap.
Donor Mood: I construct an aggregate variable by calculating the CFscore of the median first-time donor to proxy public mood.
Fundraising Success: This is measured by the log total amount raised by each candidate and the log sum of unique contributors.
Source of Funds: For each candidate, I calculate the percentages of funds raised from PACs, individual donors, party committees, and self-funding.
In my previous post, I looked at the ideology of Google’s employees and board of directors. I have since extended the analysis to board members from twenty major U.S. corporations, including the top ten corporations on the Fortune 500 list. The ideological measures include information from all contributions to state and federal candidates as well as contributions to party and ballot committees. Nearly all members of major corporate boards have made political donations in some form of another. Overall, I was able to estimate positions for about 90 percent of board members from their contributions to candidates or party committees between 1992 and 2008.
The graph below displays the ideological positions of members grouped by corporation. Each row represents a corporation and the points along the line are the locations of its board members. The point size represents the log-scaled total amount given. In addition to the board member positions, I included the median position for each corporate board (black triangles).
These results challenge conventional beliefs about the political leanings of corporate leaders. Republicans have long been seen as the party of big business. To whatever extent this label should apply, it probably owes more to the party’s policies than the composition of its support base. Although board members from some sectors exhibit conservative allegiances—notably the oil, gas, and coal industries—most corporate boards are either dispersed across the ideological spectrum, or seem to have aligned with the left, as is the case of many of the growth stories of the new economy. To provide some context, an ideal point at zero places a contributor directly at the political center, almost exactly at the midpoint between the mean Republican candidate and the mean Democratic candidate. The median ideal point is 0.12 for board members from the top ten corporations on the Forbes 500 list, and is -0.03 for a not quite random sample of 409 board members from other major corporations.
There has been much discussion about how the Supreme Courts’ recent ruling in Citizens United v. FEC will change how corporations involve themselves in elections. The estimates of corporate board members might provide insight on the matter. That being said, I have two preliminary predictions:
1) Corporate boards dominated by liberals/conservatives will be far more likely to become involved in partisan elections.
Boards composed exclusively of liberals/conservatives should have little difficulty agreeing on which side to support. Whatever disagreement does arise will center on whether allocating the funds is worth it. On the other hand, deciding which candidate to support could be a much more contentious issue for boards with both liberal and conservative members. Even if a clear majority agrees to support the Democratic/Republican candidate, forcing a vote would risk upsetting or alienating the conservative/liberal board members. It just does not make sense for a board to engage in partisan conflict when they could easily compromise on not spending the money on either candidate, or better yet, spending it on the type of non-partisan issue ads that are already common. Simply put, bi-partisan boards will rarely take part in partisan politics.
2) Republicans will not be the clear beneficiaries of Citizens United.
Although Citizens United has routinely been interpreted as a favorable development for Republicans, the ideological positions of board members suggest this assessment is premature and might prove incorrect. Although the survey of corporate boards is incomplete, liberals appear to control as many boards as conservatives do.
Insofar as my first prediction is correct, the preponderance of bi-partisan boards means that the corporate money flowing into elections will be closer to a trickle than a flood. If not, then perhaps it will be a relief that the ideological distribution of the people who control large corporations looks quite similar to the population at large.
I updated the industry and occupational rankings graph to include all repeat contributors in election cycles between 1990 and 2008. Anyone who has seen the previous figure will notice an overall shift the right for most industries and occupations. By including two-decades worth of election cycles, the estimates tend to smooth out the shift to the left caused by the unusually favorable climate for Democrats during the 2008 election. On the other hand, it tends to understate the extent to which certain industries have moved to the extremes in recent years.
I also added a few features to the plots in response to being rightly called out for the sloppy/incomplete presentation of the previous graph by an expert on statistical graphics. These new and improved graphs include horizontal bars that represent the 40-60 percentile of each industry/occupation that are intended as rough measures of dispersion–most industries are bimodal, which would make wider distribution bands span nearly the entire graph, hence losing most of their informational content. In addition, the y-axis now indicates the amount given by repeat contributors within each industry/occupation. These amounts would be much larger if non-repeat contributors were included in the sample as well.
Obviously, not all industry and occupational categories could be included in the graphs. I plan to post a spreadsheet with a complete ranking of categories sometime in the near future.
Regardless of one’s beliefs about how the 2008 Presidential Election ranks among other historic elections, it was certainly the most active and actively watched money race in the nation’s history. The $1.75 billion raised by presidential hopefuls from both parties more than doubled the amount raised during the 2004 Presidential Election. The surge of candidates declaring their intentions to run within months of the 2006 Midterm elections and the storied showdown between the Obama and Clinton campaigns for the Democratic nomination led to a primary season that started earlier and ended later than any in recent memory. All of this provided a seemingly endless stream of contribution data that, among other things, is ideal for tracking the ideological progression of the candidates over the course of the election. The figure below combines the IMWA with smoothing techniques to get an idea of how the election progressed.
Ideological Progression of Candidates during the 2008 Presidential Election: The presidential candidate trends are estimated in a two-step process. This first step recovered ideological estimates from an IMWA scaling that included all contributions from the 3125 PACs and 131,000 individuals that gave to two or more federal campaigns during the 2007-2008 Election Cycle. Holding the contributors estimates static, the presidential candidate trends are then smoothed over time using locally-weighted regression (LOESS) with a span of .3.
Although the plot largely speaks for itself, there were a few features that caught my attention. The first is that, with the exception of Biden and Richardson, the Democratic candidates generally maintain a consistent rank ordering, whereas the ordering of the Republican candidates is much more chaotic. The second point of interest is the Obama campaign’s move to the center during the run-up to the general election. I would venture to guess that this was this was the combined effect of Sarah Palin’s VP nomination and the financial meltdown, but that is just speculation.
Partisan giving rates have long been the standard measure of contributor ideology. Measured simply as the percentage of funds a contributor gives to Republicans/Democrats, the partisan giving rate is a straightforward, easy to understand and generally useful measure of contributor ideology. Although it is a great place to start, it is a far rougher approximation of ideology than most people realize.
In their book Polarized America, McCarty, Poole and Rosenthal (MPR) introduce a method of recovering the ideology of contributors that does away with the partisan assumption without introducing a handful of more complex assumptions in its place. Rather than assume all members of each party shared a single point in ideological space, they used as starting points the ideological scores of each Congress member, as reveal from Congressional voting records via their DW-NOMINATE method. This relaxes the assumption that Tom Coburn and Olympia Snowe are ideologically indistinguishable—the same being true for Dennis Kucinich and Ben Nelson, for that matter. By first ordering Congress members along the liberal/moderate/conservative spectrum, MPR can then look at the contribution profiles of each contributor and construct an ideological score based on money-weighted averages of the recipients’ ideology. For example, if a contributor gave $1000, $500, and $1500 to three Democratic candidates with the respective ideal points of -0.5, -0.1, and -0.4, the ideal point estimate given to that contributor would be:
(-0.5*1000 + -0.1*500 + -0.4*1500)/(1000+500+1500) = -0.383.
MPR’s method of estimating the ideal points of contributors is a simple yet powerful tool. Nonetheless, its reliance on DW-NOMINATE scores limits which contributors and candidates can be included in the model as well as what we can take away from the results. Notably, it excludes contributions to unsuccessful candidates that never hold office and compile a voting record.
The Iterated Money-Weighted Averaging (IMWA) procedure takes the MPR procedure a few steps further. If we can estimate the ideological positions of contributors based on the ideology of the recipient candidates, there is no reason why we cannot simply flip the procedure on its head and recover ideological estimates of candidates based on the ideology of their financial supporters. This results in a more inclusive ideological map of campaign finance that is not reliant ideological scores of candidates recovered from voting records.
The algorithm is as follows:
1) Set the ideal points of all Republicans candidates to 1, Democrats to -1, and independents to 0.
2) Estimate contributor coordinates as a function of the money weighted averages of their contributions.
3) Estimate legislator coordinates as a function of the money weighted averages of their contributors.
4) Normalize legislators to have a mean of 0 and a standard deviation of 1.
5) Go to step 2; repeat until convergence.
To illustrate, I include results from a scaling of the 2008 Election cycle that uses data from all individuals and PACs that contributed to two or more unique candidates. The figure above compare the ideological distributions of lobbyists, an important category of individual contributors, first derived from the partisan measure and then from the IMWA measure. The partisan measure lumps most lobbyists at either 0 or 1, representing the preponderance of partisans lobbyists that give exclusively to Democrats or exclusively to Republicans.
The IMWA estimates provide a much more fleshed-out portrayal that should lead to more informed inferences. For example, while lobbyists are clearly polarized, the IMWA distribution reveals that lobbyists are not as polarized as the partisan measure suggests. In fact, lobbyists are one of the least polarized industries. It also appears to be the case that liberal lobbyists are more dispersed than their conservative counterparts on K Street. In other words, the IWMA estimates are better able to account for something that is painfully obvious to anyone who follows politics, that ideology is not uniform across members of the same party or their supporters and that it is possible for one party to be more dispersed than the other.
The “no frills” IMWA method is not the only way to construct an ideological map of campaign finance. There are plenty of alternative approaches, most some slight variation on correspondence analysis, although more sophisticated statistical methods exist as well. The take-away is that no matter how one slices it, campaign finance data is rich in ideological content waiting to be unlocked and explored.