Almost a quarter of the $1.2 billion spent in this election cycle ($292.6 million.) has come from dark money organizations, typically 501(c) groups, which are not required to disclose their donors. These groups, led by Karl Rove’s Crossroads GPS and the U.S. Chamber of Commerce, are afforded such leeway because they cast themselves as “social welfare organizations” engaged in non-political issue advocacy. These organizations claim to be independent of the political party infrastructure, and the overtly political super PACs, but their behavior suggests otherwise. We find, instead, that many of these groups’ allocations of resources closely resembles the patterns observed in party committees, like the Democratic Senatorial Campaign Committee (DSCC) or the National Republican Congressional Committee (NRCC).

We looked at the 33 outside groups with over $2 million in total spending as of November 2nd, excluding those organizations, which had only spent on the presidential race. By looking at which candidates these groups have spent in support or opposition of, as well as how much they have spent, we have calculated the group’s spending similarity to the Republican and Democratic Senatorial and Congressional committees. These scores range from -1 to 1. A score of 1 indicates that a group allocates its funds across the relevant races in exactly the same ratios as the committee to which it is being compared, and -1 indicates perfectly opposed allocations. Of the groups analyzed spending in the Senate, the average similarity score for that group as compared to the party committee with which it is aligned was 0.395. In the house the effects were notably weaker, with an average similarity score as compared to the aligned party committee of only 0.171. This indicates that outside groups pick and choose the House races in which to be involved much more selectively in the House than the Senate, as compared to party committees whose spending is more widespread.

A highly similar spending allocation indicates that an organization is working towards the same strategic goals as one of the party committees, while a directly opposing allocation indicates that it is working against that committee’s goals. One might expect that non-political groups would appear more independent than the overtly partisan super PACs. Were this the case, ‘non-political’ 501(c)s would tend to have scores closer to 0 than super PACs. However, analysis of the data here finds no evidence, that the spending of these 501(c) groups is any less partisan than the registered super PACs.

Our analysis found no statistically significant variation in these similarity scores between super PACs and the 501(c)s. If anything, the dark money groups seem to emulate one party’s spending or oppose the others spending to greater extent than do the super PACs. Instead, we find high levels of similarity between the party committees and their aligned outside group, which calls in to question the independence of these controversial groups.

The bubble chart below shows the position of these top top outside groups in terms of their similarity to the DSCC and the NRSC. Size is proportional to the amount of money the group has spent on Senate candidates since September 7th (reporting became required for dark-money groups). Two coloring options are available below the chart to highlight different features. Groups can be colored either according to the party with which they are aligned or by organization type (super PACs are shown in orange, non-party non-committee in green).

In Senate race spending, the Harry Reid advised Majority PAC, mirrors the Democratic Senatorial Campaign Committee most closely with an extremely high similarity score of 0.89. Majority PAC is followed closely by the union groups Service Employees International Union Committee on Political Education (SEUI COPE), and American Federation of State County and Municipal Employees (AFSCME ) PEOPLE (Public Employees Organized to Promote Legislative Equality) with scores of 0.83 and 0.81 respectively. On the Republican supporting side of things, the pack is led by Crossroads GPS, Americans for Tax Reform and the U.S. Chamber of Commerce whose spending highly emulates activities of the National Republican Senatorial Committee, with scores of 0.81, 0.58 and 0.54. Of these six organizations, only Majority PAC – which is a registered super PAC and must disclose its donors — is permitted to engage in explicitly political activity according to its tax-status.

While groups like Majority PAC or Crossroads GPS with high similarity scores look a lot like party committees in terms of how they allocate their resources, other groups display little similarity to either party organizations’ spending, but are extremely dissimilar to one of the party committees. Focus on the Family Action, a 501(c) associated with conservative evangelical James Dobson, is a particularly clear example of this. While its spending is not particularly similar to the NRSC (.045), it is very dissimilar to the DSCC, with a score of -0.78. This indicates that rather than spending in the same manner as the Republican Party establishment, it has allocated money in response to democratic money, spending on negative ads on candidates where the DSCC is spending on positive ones and vice versa. The NRA Institute for Legislative Action, Freedomworks for America, Ending Spending Action Fund and the American Future Fund all seem to follow the same pattern as Focus on the Family Action. These conservative aligned groups’ spending appears to respond to democratic spending rather than emulate republican spending.

In contrast, this pattern appears to be rare among democratically aligned outside groups. Only SEIU – PEA Federal and FSA Pac appear to spend to oppose Republican Party money rather than supplement democratic spending.

The bubble chart below shows the position of the top outside groups in terms of their similarity to the DCCC and the NRCC. Size is proportional to the amount of money the group has spent in on House candidates, since September 7th (reporting became required for dark-money groups). Two coloring options are available below the chart to highlight different features. Groups can be colored either according to the party with which they are aligned or by organization type (super PACs are shown in orange, non-party non-committee in green).

Spending activities of these groups on House candidates reveal similar, although less pronounced patterns as Senate spending. The U.S. Chamber of Commerce, is far-and-away the outside group whose House spending activity is most similar to either party’s congressional campaign committee activity, with a similarity score of 0.61 as compared to the National Republican Congressional Committee (NRCC). House Majority PAC and SEIU COPE remain the outside groups with the most similar spending to the Democratic Party committee’s (DCCC), spending in the House as they were in the Senate.

SEIU-PEA Federal and the Planned Parenthood Action Fund are the Democratic affiliated outside groups which seem to respond most strongly to Republican Party money, while both the NRA Institute for Legislative Action and the NRA of America Political Victory Fund, are the strongest conservative responders to Democratic Party spending.

Conclusion: This analysis begins to offer insight into how we should understand the spending decisions of dark money groups. Some groups appear to primarily aid the party committees with which they are aligned, others groups seem to take their cues from the party spending of the opposition. This pattern, which appears more common among conservative affiliated organizations, takes the form of pouring opposing money into the places where democrats are spending heavily. While a time-series analysis would be necessary to more precisely uncover the causal relationship between the party committees and outside groups, this indicates a first step in understanding the various strategic intentions of these otherwise hard-to-pin-down organizations.

As far as spending patterns are concerned, there is no evident no link between an organization’s level of partisanship, and whether that group is a super PAC or a 501(c). This is not entirely surprising, as many of these dark money groups are run by the same people who run super PACs. However, it is significant evidence in support of re-evaluating the tax-exempt, non-profit status of these groups that affords them the ability to keep their donors secret. Most significantly, when taken as a whole the similarity scores, particularly among groups engaged in significant spending in the senate make a begin to make a compelling case that these outside groups are neither meaningfully non-political (in the case of 501s) nor independent.

Full similarity results for the Senate and House are displayed below:

Senate:

House:

Similarity Score Methodology:

Similarity was calculated using a fairly standard vector space model. Senate spending was represented as a multi-dimensional space composed of the 25 dimensions, each representing one of the candidates with the most spending for or against them. Each organization was represented as (including the party committees) was represented as a vector within this space, with a value for the amount that organization has spent for (positive) or against (negative) that candidate. Similarity between two groups was calculated by taking the cosine of those groups’ vectors.  Thus, the similarity score for SEIU COPE as compared to the DSCC represents the cosine of the angle between the SEIU COPE vector and the DSCC vector.

The same procedure was followed for House races. The multi-dimensional space for the house race was based on the 70 House candidates at which the most outside spending was directed.