The analysis was conducted at the request of and in collaboration with This American Life and Planet Money public radio programs. It is featured on the March 31, 2012 episode of This American Life.
When it comes to fundraising potential, not all committee assignments in the U.S. House are created equal. A new Sunlight analysis of House fundraising patterns from the 103rd to 111th Congress finds that three committees – Ways and Means, Financial Services, and Energy and Commerce – stand out as being particularly lucrative. Others – most notably Judiciary, House Administration, and Natural Resources, appear to be fundraising duds.
Table 1 shows are the estimated committee “bonuses” for individual member itemized and PAC fundraising totals. Relationships that are statistically significant are in bold.
Table 1. Estimated committee assignment “bonuses”
|Committee Assignment||Estimated itemized contributions “bonus”||Estimated PAC contribution “bonus”|
|Ways and Means||+$258,924||+$295,774|
|Energy and Commerce||+$142,030||+$197,220|
These estimates come from a multivariate regression that compares all members across multiple sessions of congress, controlling for seniority, majority status, and session of congress, with fundraising totals adjusted for inflation. The complete regression results can be found here. See below for a more complete explanation for how we came to these results.
Another way to evaluate the value of committees is to look at patterns of incumbent switching. Table 2 looks at what assignments incumbent House members selected (or were given) from the 104th through 111th Congress.
Table 2. Incumbent committee switches
|Ways and Means||53||4||13.3-to-1|
|Energy and Commerce||76||10||7.6-to-1|
Certain committees are clearly more popular than others. Over almost two decades, Ways and Means had 53 incumbents join, but only four leave. Energy and Commerce had 76 incumbents join, but only 10 leave, and Appropriations had 62 incumbents join, but only 11 leave.
Other committees have different flows. Judging by members’ behaviors, the least desirable committee to be on is Small Business – only 18 incumbents joined, while 100 left. Agriculture and Veterans’ Affairs are also committees that incumbents leave far more often than they join.
This way of ranking committees generally correlates with the fundraising prowess of the committees (Table 1). Ways and Means tops both lists, and Energy and Commerce is in the top three of both lists. However, Financial Services is a net loser among incumbents. Perhaps this is because many members find the subject matter too complicated or boring for their tastes. And Judiciary, which is associated with lower fundraising flows, is just about in the middle when it comes to the balance of incumbent leavers and joiners.
It is important to recognize that members may care about more than just raising money, and may have genuine policy interests. Also, members who are in relatively safe districts and have fewer ambitions to rise up in the party may be quite happy to serve on committees with less fundraising potential.
Methodology: How we got these results
The estimates for committee fundraising “bonuses” (Table 1) come from the results of regressions that estimate different aspects of individual member fundraising as a function of members’ committee assignments and ranking on those committees, controlling for majority party, seniority, and different congresses. We’ve adjusted for inflation so that all fundraising totals are in 2012 dollars.
The complete results of this series of regressions are published here: https://data.sunlightlabs.com/dataset/Regression-1-Effects-Of-Committee-Assignments-On-C/dk8x-utgd
The results report four different sets of columns. The first set of columns estimates the effect of committee assignments on the total itemized contributions. The second set of columns estimates the effect on individual itemized contributions. The third column estimates the effect on PAC contributions. The fourth estimates itemized out-of-state contributions. Essentially, what we have done is compared all members to all other members across multiple Congresses, trying to isolate the statistical relationship of committee assignments and fundraising totals.
For each set of columns, the predicted effect of different assignments and control variables are in the first column, the standard error in the second, and the t-stat (the effect divided by the standard error, a measure of statistical significance) in the third.
So, for example, being on the Ways and Means committee is associated with $258,924 more in total itemized contributions, as compared to not being on the committee, all else being equal. The standard error for this estimate is $63,543, so we can think of $258,924 +/- $63,543 as being the predicted fundraising advantage of a seat on the Ways and Means committee. The t-statistic is 4.07, which means that it’s a statistically significant relationship. Statisticians use 1.96 as the cut-off low point for clearly significant statistical relationships. Generally, a t-stat of 1.64 is considered the lower bound for statistical significance.
The row at the bottom, “adjusted R-squared” is a measure of how much variation each model explains. The regression estimating total itemized contributions explains 20.8% of the variation across members. This means that about 79% of the variation across members is explained by factors we didn’t include in the model (such as unique member and district characteristics). The higher the R-squared, the more variation we’ve explained. The lower the R-squared, the more variables we are probably missing.
If we wanted to estimate predicted fundraising for a particular member, we first would get the baseline fundraising for each category for each member by starting at the intercept. Then we add an additional bonus for the congress we are in, multiply the number of years in the chamber by the value in the seniority row, and add an extra bonus for the majority row. Then we add the value for each committee assignment. If we are looking at a chair or a ranking member, we add an extra bonus on top of the already predicted value for being on the committee.
So, for example, if we had a fourth-term Democratic member in the 111th Congress serving on the Ways and Means Committee, we could estimate predicted PAC contributions by starting with the intercept of $309,723, adding an extra $266,056 for the 111th Congress, less $28,834 ($7,096 * 4 terms) for seniority + $66,877 for being in the majority. So we are at $613,3822 in PAC contributions baseline. Now we’d add an extra $295,774 for being on Ways and Means. So the predicted PAC fundraising total would be $909,596.
Now the caveats.
These numbers measure statistical patterns, but they don’t directly show causality. These numbers tell us that members on certain committees raise more or less than their peers on other committees. But it may be the case that better fundraisers get rewarded with better committee assignments. As always, one must be careful confusing correlation with causality.
Still, there are some good reasons to infer some causal relations. One big reason is that the fundraising bonuses on the big money committees come primarily in the form of PAC contributions, which mostly come from corporations and are more likely to have committee-specific interests. Both Ways and Means and Energy and Commerce are associated with less individual fundraising than your average member. Financial Services is associated with more individual giving, which makes sense since individuals who work in the financial sector are major donors. But still, two-thirds of the Financial Services committee “bonus” comes from PAC contributions. Also, all three committees are associated with significantly more out-of-state funding than your average member.
Another way to get at causality is to estimate a time series regression, where we explain member fundraising as a function of prior fundraising efforts (controlling again for majority status, seniority, congress), and then look at the effect of committee switches on fundraising.
The results of these estimates can be found here: https://data.sunlightlabs.com/dataset/Regression-2-Effects-Of-Changing-Committees-On-Can/j3sk-hyqf
These estimates do a better job of explaining variation across members (notice the higher R-squared values) because they effectively take into account the differences across members, including their district characteristics and fundraising prowess.
The only statistically significant results for changing committee assignments involve transferring to Ways and Means, Financial Services, and Energy and Commerce.
Getting on Ways and Means boosts your PAC fundraising levels by $208,315, your out-of-state itemized fundraising totals by $233,742, and your total itemized fundraising totals by $220,598, all in statistically significant ways.
Joining Financial Services produces boost of $101,695 to PAC contributions and a boost of $140,334 to your out-of-state contributions, both statistically significant.
Joining Energy and Commerce yields a statistically significant boost of $72,933 to your PAC contributions.
These results should give us even more confidence that Ways and Means, Financial Services, and Energy and Commerce provide genuine fundraising bonuses to members.
There are probably two reasons why we don’t see any statistically significant results for the effect of joining a dud fundraising committee.
One is that there are only a limited number of cases where incumbent members switch onto the dud fundraising committees, and with only a limited number of cases, it is hard to get statistical significance because it is harder to distinguish patterns from chance.
The second reason is that if everybody knows which committees are dud fundraising committees, those who join those committees are members who are probably less likely to be concerned about fundraising, perhaps because they have safe seats. Moreover, if members are joining a dud fundraising committee, it’s more likely to be from another dud fundraising committee. In this case, it’s less likely to alter fundraising.
Overall, these results confirm that some House committees are better for member fundraising than others. In particular, Ways and Means, Financial Services, and Energy and Commerce are very good fundraising committees. And for good reason: Ways and Means has jurisdiction over tax policy, Financial Services over securities and banking policy, and Energy and Commerce over energy policy. In all three policy areas, a substantial number of corporations care very much about how policy gets made, and their employees are willing to contribute substantial sums – both through their PACs and individually – to make sure that they have access.