Leadership in the Senate: New Boss Same as the Old Boss?

The 2024 election has clearly injected real novelty into American politics. It’s not every plebiscite that sees a billionaire, with much help from other billionaire backers, win an electoral plurality and then turn over key policy moves to an even richer billionaire. Or that shuttles so many wealthy industry chiefs into cabinet positions. But vast as the powers of the executive branch are, the fate of the Trump agenda will still depend crucially on the Republican leaders in Congress marshalling votes to pass the President’s Christmas list of tax abatement, program cuts, and other sweeping policy changes with historic distributional implications.

It thus makes sense to look away from the sound and fury over DOGE and scrutinize the political finances of the leaders of the new Congress. The full shape of 2024’s political money is still being tabulated. But the two previous presidential elections of 2016 and 2020 offer plenty of insights into what researchers are likely to discover. In both, MAGA was underwritten by a singular mix of big donor financing blended with small-donor enthusiasm for the WWF-style political brawling that Trump excelled at. Ferguson, Jorgensen, and Chen styled the bipolar curve a “barbell”: big-donor checks in the hundreds of thousands on one side versus a lesser swarm of small-donor contributions. Still, this Trump donor profile differed markedly from that of the other key players in DC Central: the congressional leaders of both parties and the other presidential candidates all surfed on the foam of large-donor largesse pretty much exclusively. With one exception, of course: Bernie Sanders, who received virtually no support from big donors and relied almost entirely on donor contributions of less than $200.

Mitch McConnell, who just stepped down as Senate Republican leader, perfectly exemplified this other spectrum with a near-total dominance of big donors. And his successor in the Senate’s top spot, John Thune of South Dakota, looks set to replicate the pattern.

In an earlier review of McConnell’s political money, we looked at the outgoing Senate Leader’s campaign finances in detail and concluded that what distinguished him was his pivotal slot within a web of big money and K-Street (lobbying) interests. Now that McConnell has stepped aside, it is time to ask if the “generational change” in the Senate many are clamoring for is likely to yield much apart from more of the same.

In his remarks upon winning the caucus’s vote to replace McConnell as Republican Leader, Thune sounded as though he intended to herald a dramatic departure: He proclaimed “a new day in the United States Senate, a new day in America”.

But the phrase could, of course, also mean “a day like any other day”, and this in fact is the only reasonable way to describe the new order.

Figure 1 summarizes the situation. It compares two-year snapshots of the campaign finances of Thune and McConnell, covering the 2022 and 2024 cycles. The key observation is hard to miss: starting from a fairly typical distribution of contributors centered in the $5,000-$10,000 range, we find that even within standard campaign finance limits, the working class donor—who’s not cutting $1,000+ checks to political committees—barely shows.

As he ascended to the leadership, Thune’s financing underwent a fairly abrupt change-of-state: the 2023-2024 contributions came overwhelmingly from the megadonors (purple bars to the right).

What happened exactly? Thune’s 2024 efforts included starting up two new entities: a paired Super PAC and a dark money group, both called New Heights for America (NHA). The Super PAC can receive checks of any amount from any entity (individuals and corporations included), with the amounts it spends likewise unlimited.[2] By tradition, congressional leaders maintain such Super PACs, led by their former staff, which raise hundreds of millions to deploy in battleground districts and to provide the respective leader with a golden whip to control their chamber.

It is worth noting that even prior to 2024, Thune’s small donor receipts were almost derisory: 7% across all of his PACs. With the new Senate Leader’s Super PAC in play, that number shrinks to less than 2%. In 2025, we expect, the transformation will be complete, with Thune an heir to his progenitor’s monumental money machine.

It is worth delving into the mechanics of this succession, which could have important implications for how various Senate battles play out. On our reading, control of McConnell’s money is contested turf. The key piece on the board is the Senate Leadership Fund (SLF)—consistently the Senate’s biggest Super PAC—which until December was controlled by ex-McConnell chief of staff Steven Law. Following McConnell’s departure, a new leader was appointed in a selection process that Thune political staffers headed, with SLF now co-directed by former GOP Senator Cory Gardner (R-CO, formerly a Thune deputy whip) and Trump Special Assistant Alex Latcham, a dyad that likely reflects the Trump administration’s desire for a more direct role in setting Senate priorities, given Trump and McConnell’s notorious divergence on Senate candidates in the 2022 midterms.[3] McConnell’s dark money group One Nation, the largest spender of dark money from undisclosed sources in 2020 and 2022, seems to have been put into more Trump-aligned hands, going off the evidence of Latcham’s Twitter bio—the only source we could find identifying who is in charge of this hundred-million-dollar organization, One Nation’s website being silent on the matter.[4]

The gap between McConnell’s $300 million baton and the approximately $11 million raised by Thune’s pop-up Super PAC, New Heights for America, may therefore signify a new power-sharing arrangement between the Senate leadership and the Trump administration—a somewhat surprising and perhaps fragile formation given the proverbial shellacking that GOP candidates took in the 2022 midterms, which media reports widely attributed to Trump.

This new arrangement may well have implications for Senate policymaking and could become a point of conflict in the future. Conflict, to be sure, of a relatively minor sort; the consensus on such issues as eliminating Biden-era climate regulations, mass layoffs in government, or reauthorizing the top-heavy Trump tax cuts, seems rock-solid.

But not on every issue. Dissent has rumbled, for example, on the Trump tariffs, especially from farming constituencies fearing retaliation against their own exports. Thune himself told Politico, “I’m not a big fan of tariffs, and I made that clear during the last Trump administration for the reason you just mentioned, because the retaliatory action taken by countries with whom we need to do business and who are big markets for American agriculture.” Biden-era industrial policies were naturally popular with big firms like Intel and TSMC receiving incentives intended to re-shore semiconductor manufacturing, and Thune is said to have been wholly blindsided by Trump’s announcement that he seeks to cancel CHIPs Act subsidies. MAGA-style working class carve-outs like no taxes on tips, Social Security, and overtime will require (at least apparent) revenue offsets, and the President has revived his 2016 campaign promise to squeeze some of that cash from private equity fund managers by eliminating the carried interest loophole. The private equity industry, which spends lavishly on both Democrats and Republicans, is predictably not thrilled.

The subject of carried interest is worth a few paragraphs more since the issue hits at some core disputes between the Reaganite wing of the party (including Thune and McConnell) on the one hand and Trump on the other. The loophole allows money managers holding an asset on an investor’s behalf to benefit from lower capital gains rates on their share of profits from the investment, that rate being about half the ordinary income rate at the upper end. The concept of taxing bankers touches a nerve of the post-Occupy Wall Street populism that Trump has appealed to —despite the fact that its actual revenue effects would be stunningly mild,[5] serving as an offset more for the passions of the public than the federal balance sheet.

Obama pledged to end the loophole in his 2008 campaign and included provisions to do so in several of his budget proposals, which were scissored to death by lobbyists and legislators. Trump made it a wedge issue in 2016 debates (though Clinton was formally in favor of closing the loophole), railing in a 2015 interview—with all the low-quant Old Money swagger of a Forbes Billionaire with a 4/10 self-made score (“fortune inherited”)—that hedge fund managers were “guys that shift paper around” who were “getting away with murder”.[6] Closing the loophole entered the draft version of the 2017 tax cuts bill and fell to lobbyists’ blades with at most tepid opposition from Trump’s Treasury secretary Steven Mnuchin, a Goldman-Sachs alum. The final bill included a much diluted provision restricting enjoyment of the provision to assets held for more than three years, which as a Congressional Research Service study noted would barely affect private equity, which typically holds assets for longer than five years (hedge funds tend to trade at higher frequency and are taxed at the ordinary rate).

Biden followed the Obama protocol with budgets that attempted to close the loophole in full, and Trump has again repeated the 2016 promise, whose defects as “serious economic populism” should be obvious from our discussion. So, with carried interest again on the table, MAGA’s ability to deliver even this concession to working class voters will again be put to the test. Industry players are gearing up for a familiar routine: to “go up there and educate”, as Venture Capital Association President and CEO Bobby Franklin was quoted in Politico. Meaning: to Capitol Hill, to lobby.

Financiers are reported to be hopeful that Thune will serve as their firewall to any assault on carried interest.[7] And they have good reason to do so. For, during their meetings to “educate” the new Republican leader, the lobbyists who swarm Capitol Hill year after year will be able to share moments of water-cooler camaraderie with a Senate leader who himself has the distinction of having played that role too.

For Thune himself incarnates the revolving door, having worked as a lobbyist with a concentration in rail. He started in politics as South Dakota’s Republican party boss (1989-1991), before receiving a gubernatorial appointment as the state’s chief railroad regulator. After a three-term stint in the House, Thune exited through the revolving door, registering as a lobbyist and using his political connections on behalf of the Dakota, Minnesota and Eastern Railway (DME)—which operated the railways he previously regulated—racking up $220,000 in lobbying fees from that client as the Thune Group (a group of one). Ascending to the Senate in 2005, the freshman member of the Environment & Public Works Committee[8] authored provisions inserted into the Railroad Rehabilitation and Financing Innovation Act (RRFI) reauthorization, provisions that multiplied by ten the amount of federal loan financing allocated for railroad infrastructure projects and with billions set aside for regional railways, including his former client.[9]

The suspicion that Thune may be receptive to lobbyists’ pleas is strengthened by the strong similarities between his revolving door network and McConnell’s. In Figure 2, we have plotted the firms that employed lobbyists affiliated with either of these two Senators, meaning staffers who have previously worked for Capitol Hill offices and have disclosed links to McConnell or Thune. While some Senators in our lobbying dataset have affiliates who go on to lobby (in rare cases exclusively) for labor or nonindustrial advocacy groups, what leaps out from inspecting the two Senate GOP leaders is: industry, industry, and more industry, particularly of the “big business” variety. Among the sectors who’ve found footholds in both these highest of legislative offices, we find wireless oligopolies, pharma interests (Novo Nordisk, Boehringer, Johnson & Johnson), the energy sector (American Petroleum Institute), real estate (National Multifamily Housing Association, Real Estate Roundtable), and defense (Honeywell). Regarding the carried interest fight, the financial industry’s clutch upon the Senate seems about as secure as Fort Knox. Among the list of companies deploying Thune-affiliated lobbyists, we find ample representation from the big names of finance: JP Morgan, Goldman Sachs, the American Bankers Association, and the American Investment Council—private equity’s lobbying arm, which spent some $20 million lobbying Congress since 2016 and employs former Thune Senior Advisor Douglas Schwartz to lobby on “tax issues affecting private equity and partnership” (aka, carried interest).

A final set of data on similarities between Thune and McConnell comes from their voting records.[10] Congressional roll calls provide thousands of observations of legislators’ preferences per year, and is probably the most data-dense environment in which to study legislator behavior, since most Senators weigh in on most issues that come before the Senate—though obviously the set of bills brought to the floor is pre-selected and hardly spans the space of policy alternatives. Nor does it testify about behavior inside closed committee discussions.

How do the new GOP leader’s votes compare with those of his predecessor? We have measured this in two ways, shown in Figure 3. The top panel plots the 118^th Senate in relation to a statistical construct from Poole and Rosenthal, based on the “ideal point” model of legislative preferences that aims to compactly represent legislator ideology on two axes by fitting a numerical model that linearly predicts most votes in that constrained landscape.[11] We analyze Axis 1, which tracks legislators’ positions on the basic Left-Right spectrum, since the interpretation of Axis 2 has been widely questioned, and its explanatory power is marginal.[12] The second panel is provided to show robustness: we count the proportion of times each Senator’s votes agreed with the common anchor point of McConnell, a measure that is more directly interpretable and more proximate to the data.

As the Republican leadership changes hands, its (pro-big business) center clearly holds. Thune (and we note, Texas Senator Jon Cornyn, another contender for the Senate’s top slot as well) are of a feather with McConnell, with Thune sitting at number 9 in proximity to the former leader on both measures. Thune agreed with McConnell on 89% of votes, putting him in the top decile of Senators in vote agreement with the former majority leader. By contrast, Rick Scott, a favorite of some big donors, including notably Elon Musk, was knocked out in the first ballot, which does not surprise us given his more center-right positioning by the roll call measures.

What are the determinants of a Senator’s roll call record? The individual personality is one, surely. The constituency and their policy preferences another. We focus on the donors: who finances the Senator’s political activities? In Figure 4 (x-axis), we have combined contributions from political action committees (PACs) and individual contributors to each Senator’s campaign committees and leadership PACs,[13] grouping contributions by the donor (including individuals who contributed to multiple candidates). The output is a similarity measure: a one-dimensional index of the extent to which two politicians’ campaign money comes from the same sources. (See the footnote for technical details.[14])

By this measure of donor similarity, Thune happens to be McConnell’s nearest neighbor even prior to stepping into the latter’s structural role. Expectations should, therefore, be set that whatever differences might exist between the two for now, the force of donor influence is likely to bring them into even greater alignment—as the qualitative data reviewed above would tend to indicate as well.

This leads to a final question. We have two empirical bases for inspecting how Senators cluster: on the one hand the donors, and on the other the votes. Do legislators who vote similarly also have a similar donor base?

Birds of a feather flock together, or as we put it in an INET working paper where the method was developed for this setting: “those who vote together get paid together also.” In Figure 4, we plot this experiment McConnell-wise, comparing each Senator’s financial similarity to McConnell with the extent to which their votes align with his. On the x-axis, we have the measure of donor-wise similarity. The y-axis presents the two indices of roll-call similarity. The analytical principles are straightforward: compare like-to-like and query how strongly the two measures relate: proximity-by-the-donors and proximity-by-the-votes.[15] The result is not surprising: as the donors flock, so do the Senators.

This point should not be slighted: money rules politics. It predicts which candidates get elected (by the magnitude of their spending; see Ferguson et. al. 2022), and as we show here, it also predicts how they vote once in office (by the extent to which their donors overlap).

Voters seem to take this for granted and very much dislike it, yet the system of lobbying, campaign finance and big donor influence just expands. According to Pew polling from 2023, Americans believe that big donors (80%), lobbyists (73%), and large employers within their districts (61%) have too much influence on members of Congress. What of the voters in their districts—you know, the people who elect them? 70% think they have too little power. “Democracy” means “rule of the people”, and the people are very much under the impression that it is not they who do the ruling.

We’ve put the focus on a comparison with McConnell in order to emphasize the continuity of the new Senate leadership with the political and financial formula that McConnell carved out during his 18-year term as Republican leader. Lest one think, though, that we are giving the Democrats a pass given that the result reported in Figure 4 is marginal for them in a statistical inference sense (p>.05), we also post below (Table 1) the result of a statistical model with fixed and variable effects for the influence of donor similarity on votes.[16] The details are explained in the note, but what is important is that no matter how you slice the data, whether we are comparing Dems-to-Dems, Reps-to-Reps, or tracing how Senators shift to vote with the other party, the influence of money is significant in every partition—and the result is robust to different measures of legislators’ roll call alignment.

If one were to place bets, a very good one is that the famous Golden Rule will persist: to understand politics in America, follow the money. When we do, we find good cause to expect McConnell’s shadow to live long beyond his tenure.

Notes

[1] Aggregated donor-wise contributions were transformed to log10 space and then segmented into 25 bins of equal log-width, with the contents of each bin summed to obtain the bar height for the bin. Note that the large bar in the rightmost bin of McConnell/2022 corresponds to aggregated contributions for more than $100 million from One Nation, a dark money group whose receipts (we infer from its spending) regularly stack up to between $75 million to more than $100 million per cycle, with sources unknown. The contributions could of course be disaggregated to the individuals and organizations that contributed to the group. The amounts would be immense here as well—reportedly with $1 million from Southern Company, $2M from Andeavor, $1M from ConocoPhillips and $1M from the American Petroleum Institute across several years (all big oil), to name some SLF donors whose identities we know only because of the diligent work of journalists. These one-off donor identifications, however, are not official tabulations, so representing these clandestine donors as a unified block is regrettably forced upon us due to this gap in campaign finance law. The same goes for New Heights for America, the Thune dark money group.

[2] NHA remains on the FEC’s books for the 2026 cycle, so an open question of some interest is the extent to which Thune’s fundraising will depend on SLF vs. NHA. Of the three contenders for the post of GOP leader, only Thune emulated the McConnell strategy of popping up a personal Super PAC in 2024. John Cornyn, who has previously used Super PACs to get himself elected (which is a different function than the battle map PACs that a Senate leader uses to mold the roster and amplify his/her influence) used a more constrained fundraising vehicle—a joint fundraising committee called Cornyn Victory (C00770180)—raising about $20 million, a quarter of which was handed to the National Republican Senatorial Committee and the rest parceled out to GOP candidates like Bernie Moreno and Larry Hogan, as well as Cornyn allies in the Senate. Checks made out to instruments such as this are restricted by the type and the number of participants, but the amounts are still astronomical ($694,000 from Sarah Perot, for example). Rick Scott, another contender for the Senate leadership, likewise enjoyed a sizable fundraising power-base as chair of the National Republican Senatorial Committee. The ability to project influence through the donors’ mighty dollars appears to be a job requirement for the Senate’s top spot, and each contender modeled it to some extent, with Thune’s case representing a fairly abrupt morph into the form of political receptacle for the big-bucks. As mentioned in the text, future control of the Senate Leadership Fund seems contested, though in any scenario, hundreds of millions will flow through channels controlled at least in part by the Senate majority leader. How various policy divergences between the Trump and the McConnell/Thune wings play out may depend on NHA’s balance sheet.

[3] Media reports abound about disagreements during the primary season, notably on Trump picks Herschel Walker and Mehmet Oz, with McConnell publicly alluding to Oz’s deficiencies in the “candidate quality” department and favoring Perdue and Loeffler for a second crack at the Georgia seat. Additional disagreements have arisen since, with McConnell running candidate-recruitment operations independently of Trump. A likely outcome of power-sharing with Trump allies co-heading the SLF is an abatement of such disputes on one side or the other.

[4] As of 3.11.2025, Latcham’s Twitter reads as follows: “Executive Director @Senate_Fund & @OneNationorg. ‘24 @TeamTrump campaign vet. Iowa native and @WhiteHouse45 alum. Father to Henry & Nora, dad jokes inevitable.”

[5] As opposed to warring over the issue, much more fundamental, of the capital-labor split in whole as regards taxation, intimations of which are scarce in the current context.

[6] As noted in a 2014 CRS report on hedge fund and private equity manager taxation, “Hedge fund trading is not always based on short-term strategies, but in general their investment horizons are shorter than those of private equity funds, whose holding periods average 6 to 10 years” (Maples 2014, CRS report RS22689). Knoll (2007) notes similarly in a legal analysis that “Because the typical hedge fund trades regularly, it generates short-term capital gain, which is taxed at the same rate as ordinary income. Thus, the tax issues raised by the carry with private equity funds are different from those raised with hedge funds.”

Numbers are hard to pin down, but recent commentators have made the same point: hedge funds are not the real beneficiary. Notably, some large funds don’t or haven’t, until recently, included carried interest (proportion-of-profits) compensation for fund managers, Larry Fink being a rather notable example. Thus, Trump’s comments leveling the charge of financial “murder” at hedge funds specifically reflect either his own ignorance or a knowing attempt to dupe the voter into thinking he was targeting more of Wall Street than he was in fact.

[7] The Politico analysis cited previously spells out the private equity industry’s playbook for this iteration of carried interest negotiations: “For now, industry officials are hopeful that Republican leaders like Majority Leader John Thune of South Dakota and Senate Finance Chair Mike Crapo of Idaho will continue to serve as a firewall against any push to treat carried interest as ordinary income. But newer members, particularly in the House, may not place as much of a premium on the private equity’s interests as their predecessors.”

[8] Charles Stewart III and Jonathan Woon. Congressional Committee Assignments Dataset, 103rd to 114th Congresses, 1993-2017: Senate. https://web.mit.edu/17.251/www/data_page.html

[9] In a contemporaneous press release announcing the rail funding measures he authored, Thune tellingly touted his “private sector” experience in the rail industry: “I advocated for this initiative while in the House of Representatives, while I was in the private sector and have continued to fight for it while serving in the Senate.” The private sector experience referred to is lobbying.

[10] Lewis, Jeffrey B., Keith Poole, Howard Rosenthal, Adam Boche, Aaron Rudkin, and Luke Sonnet (2021). Voteview: Congressional Roll-Call Votes Database. https://voteview.com/

[11] “Prediction” here is in intended in an in-sample sense: how much information about legislators can be encoded in two axes without information loss in the form of misclassified votes (for VoteView’s NOMINATE model of the 118^th Congress, the answer is about 94%). This is an important point: Roll-call embeddings are fit to the full set of data and do not issue “predictions” in the common sense.

Note: the accuracy of 2d vote classification we calculated from the VoteView `Members’ Votes` dataset for the 118^th Congress, by averaging [ p >= .5 ] for p in the column `prob` of that dataset, which indicates the model’s estimate of the probability of the vote as cast https://voteview.com/articles/…

[12] Nokken and Poole themselves note that the second axis’s explanatory power has systematically dwindled since the 1960s, prior to which it explained within-party cleavages along issues like foreign policy and civil rights, e.g., dividing Northern and Southern Democrats (see Poole & Rosenthal 1997, Ch. 6 for some examples). On the decay: “[F]ollowing the divisive battles over civil rights in the 1960s, the explanatory power of the second NOMINATE dimension has declined noticeably. […] The combination of increased ideological polarization and the evaporation of the second dimension suggest that party defectors have no choice but to shift significantly along the first dimension” (Nokken & Poole 2004). Nokken, T. P., & Poole, K. T. (2004). Congressional party defection in American history. Legislative Studies Quarterly, 29(4), 545–568. https://doi.org/10.3162/036298…

Poole, Keith T., and Rosenthal, Howard. (1997). Congress: A Political-Economic History of Roll Call Voting. Oxford University Press.

[13] Candidate-affiliated Super PACs are left out of this analysis since we cannot guarantee a comprehensive set. Their linkages to candidates are veiled by design and, as such, are not tracked by the FEC.

[14] Specifically, we use the absolute difference of (donor-wise) proportions, aka the L1 norm of the raw spend-by-donor vector divided by the total spending on that legislator. Thus, each component of a legislator vector represents that donor’s proportion of money spent on that legislator by all donors. The distance measure is bounded within [-2,2], so we convert it to a similarity by dividing the distance by 2 and subtracting from 1: finsim(i,j) = 1 – ½ * | x_i – x_j |, where x_i, x_j denote the proportion vectors for Senators i and j.

[15] One appeal of this approach is that it allows one to sidestep the process of formulating donor-by-donor and vote-by-vote hypotheses in favor of an omnibus statistical test of the extent to which donors are predictive of legislative behavior, at the cost of detailed knowledge of which donors drive the effect for which votes. That’s a lot of important detail that matters quite a lot from a practical perspective—e.g., if one wants to understand the detailed predictors of congressional policy actions that will affect real lives—and we make no bones about that, or the fact that the voting record surrenders to effective analysis within other paradigms. The approach should, however, allow one to quickly dispense with the objection that campaign contributions don’t or won’t affect how a candidate will vote (in which case, the donors are fantastically irrational).

[16] See the earlier INET working paper developing this method (Lalisse 2022) for details on the SameState controls. Our discussion here covers a four-way partition of money-effects: Republican-to-Republican, Republican-to-Democrat, Democrat-to-Democrat, and Democrat-to-Republican, which each partition testing the hypothesis: does similarity of the donor base predict similarity of the roll call. We have also controlled for the influence of two Senators being from the same state, incorporating some coarse geographic/constituency information.

Regarding the use of a mixed-effects approach, the reason for it is best illustrated graphically, as we do in the regression forest below (Figure 5). Each regression line corresponds to a linear best-fit for the effect of PAC and Individual donor financial similarity (controlling for the Senator) on roll call distances, separated by group (within- and between-party). There are clearly at least three different macro-distributions: within-party similarities for Democrats and Republicans (red and blue), and cross-party similarities (purple). In general, there is greater spread among Republicans in terms of voting records in the 118^th Congress, which we attribute to the fact that Democrats held a majority of 0 votes (1 if the Vice President is required to show up to settle a tie), meaning that passing items on the party agenda required unanimity among Democratic Senators. It follows that, for example, given low information about how members of the GOP would vote, the majority leader and majority whip would act, either by supplying bills and votes enjoying prior unanimity, or to enforce unanimity within the cohort on the bills that are brought to floor votes, leading Democrats more reliably to vote as a bloc. With a larger majority, there is wiggle room for individual Senators to dissent without obstructing the main party line and thus eliciting punishment from the whip (one need not go so far as to imagine salacious blackmail à là House of Cards: the party’s ability to direct spending should be enough to have this effect. These comments are intended as suggestions, and there is, of course, a simple way to test them: the party in power, if its majority is narrow, should correspondingly see the spread of voting profiles narrowed as well. In any case, Fig. 5 illustrates the heterogeneity between the four partitions, as well as the within-partition heterogeneity of baseline vote accordance levels (intercept term) and magnitude of the targeted effect (slope with respect to contribution similarity) is why we adopt the mixed-effects modeling approach.

[17] The table shows the results of a mixed-effects effects linear regression of the influence of PAC- and individual/organizational donor similarity (collapsed to 1d PAC&Indiv_sim) on pairwise Nokken-Poole-D1 distance between Senators. We run the same specification on each party-wise partition of the similarity matrix: Republican-to-Republican, Republican-to-Democrat, Democrat-to-Democrat, and Democrat-to-Republican, predicting within-party and cross-party vote alignment from the alignment of each legislator pair’s donors. Fixed-effect controls for SameState consist of a binary matrix with 1 if the two Senators compared share a state, and 0 if they do not. Whereas state is significant/marginal in only two (D-to-D and R-to-D), the coefficient estimates for Senator finances are significant in each partition. Regarding the effect direction, note that coefficients are negative for the simple reason that we are comparing a distance (Nokken-Poole) to a similarity (Footnote [14]). Specification: for Senators i, j we model VOTE_DIST_ij ~ INTERCEPT_fe + INTERCEPT_re(i) + β_fe^{PAC&INDIV_SIM} * PAC&INDIV_SIM_ij + β_re^{PAC&INDIV_SIM} (i) * PAC&INDIV_SIM_ij + β_fe^SAME_STATE * SAME_STATE + ε_ij. fe and re denoting fixed and random effects respectively. These results are robust to swapping the roll call measure (Nokken-Poole vs raw vote agreement) and inclusion of Nokken-Poole axis 2 as well.