Money in Politics – Greatly Overestimated and Misunderstood

The old classic about confusing cause and effect is the story of the rooster who bragged that the sun came up every morning just to hear his crowing. Money in politics is much the same — we see rivers of money flowing to the winners, and we think the winners must have won due to the rivers of money . . . not recognizing that the rivers go to the candidates who are likely to win whether any money is spent or not. This insight is one of the most persistently verified and yet overwhelmingly ignored truths of politics. Reformers love to point at campaign finance as “THE” cause of our problems, not recognizing that the design deficiencies of our election system — the way we design districts and elect a single representative from each — are the much deeper causes that gives money whatever (limited) influence it has over who makes it into the halls of power.

The Limited Impact of Money on Incumbent Success in the 2016 U.S. House General Elections

For decades, many reformers have viewed the role of money in elections as the greatest threat to American democracy. In the wake of the Supreme Court’s ruling in Citizens United, this belief has become even more pervasive. [1] While there are many ways that money might have a corrosive impact on the political process — such as buying access with legislators, funding primary challengers or influencing policy decisions — data from 2016 U.S. House races clearly suggest that campaign spending was not a decisive factor in the outcome of the vast majority of U.S. House general election races.

  • The electoral impact of campaign spending in general elections is limited No more than a few percentage points separate candidates spending millions of dollars and those spending only tens of thousands.

 

  • Spending by incumbents had no discernable impact on election outcomes. The results of our analysis suggest that only spending by challengers has a statistically significant relationship with their final share of the vote.

 

  • Challenger spending has a weak relationship with election outcomes. While statistically significant, challenger spending was associated with a smaller effect on candidate vote shares than any of the other significant variables we examined.

 

  • Money matters more in competitive races. However, candidate spending was still less impactful than other factors like district partisanship, party affiliation, and incumbents’ past performance.

As is explained in Chapter Three, “The Predictive Power of Partisanship,” fewer than 20 percent of U.S. House districts have been potentially competitive in recent elections. The hardening partisanship of voters and districts means that victory for one of the major party nominees in the remaining 80+ percent of House races is effectively out of reach, regardless of the amount of money they raise and spend. Even in competitive races, the impact of campaign spending, by candidates’ campaign committees, party committees, and independent outside spending, appears to be limited, with millions of dollars required for a candidate to buy anything more than a few percentage points in their final share of the vote.

Ultimately, the limited impact of campaign spending and the scarcity of competition in U.S. House elections mean that money was likely the decisive factor in only a handful of contests in 2016. Focusing solely on campaign spending overlooks inherent structural problems in our current winner-take-all electoral system that prevent competitive and representative U.S. House election outcomes.

Most of the academic work on the role of campaign spending in congressional elections comes from the 1990s or earlier, with authors generally concluding that money has a modest impact on outcomes,[2] but disagreeing over whether the utility of spending extends to both challengers and incumbents, or to challengers alone.[3] Others have suggested that even the modest effects attributed to campaign spending are a result of methodological problems, and that the true impact of campaign spending is in fact “extremely small.”[4]

Whatever the impact of money on elections in the 1980s and 1990s, there is reason to believe that any advantage conferred by campaign spending in U.S. House elections might be smaller today than in the past. Over the last two decades, the number of competitive U.S. House elections has continued to decline, as voters today are far less likely either to “split” their tickets by voting for candidates from more than one party or to change their vote from one election to the next.[5] With the increasing consistency of voters’ partisan preferences, fewer voters remain undecided as candidates mount their campaigns, and fewer voters are persuaded by the advertising and activity paid for with campaign funds.

Measuring the Impact of Campaign Spending in General Elections

To provide an updated assessment of the impact of campaign spending on election outcomes, we examine data from U.S. House races in 2016 and employ a regression analysis with variables measuring candidates’ spending (reported by the Center for Responsive Politics) and a range of other factors that can be expected to impact the outcomes of elections.[6] Regression analysis can provide insights about the relative importance of these different factors, and can tell us if spending is still correlated with electoral success after taking these other possible explanations into account. Note that we exclude open seats from our model.

Our analysis suggests that campaign spending by incumbent candidates is not statistically significantly associated with their final share of the vote (Figures 4.1 and 4.2).

While this finding may seem counterintuitive, it is consistent with much of the previous research on the association between campaign spending and success in U.S. House elections.

Critics of this conclusion might argue that there are methodological issues that disguise the effects of incumbent spending on their vote share. For example, incumbents’ decisions about campaign spending are likely to be influenced by perceptions of their prospects for reelection. As a result, incumbents in tighter races may spend more than those who expect an easy victory, giving the appearance of a negative correlation between spending and incumbent performance, even if in reality its impact is positive. To control for this, our model includes a variable measuring candidates’ campaign expenditures as a proportion of funds raised, on the assumption that incumbents who view their reelection as threatened will likely spend more of their campaign funds on hand (in addition to other control variables, like district partisanship and incumbents’ past performance, that are themselves highly predictive of election outcomes).

The factors analyzed in our model:

  1. Total incumbent spending ($) (including spending by outside groups)
  2. Total challenger spending ($) (including spending by outside groups)
  3. Percent of funds (%) spent by incumbent
  4. Percent of funds (%) spent by challenger
  5. District partisanship
  6. Party of the incumbent
  7. Performance over partisanship of the incumbent in 2014 and in 2012
  8. Incumbent years in office

 

Figure 4.1
Estimated Relationships with Incumbent Vote Share, 2016 U.S. House General Elections (FairVote)

 

 

 

 

Challenger spending, on the other hand, does appear to play a significant role. Still, as Figures 4.1 and and 4.2 illustrate, the magnitude of this relationship is smaller than those between incumbent vote share and the other statistically significant variables in the model for all races, and is not completely determinative of outcomes even in the most competitive races. For ease of comparison, the results for the three continuous variables in Figure 4.1 are presented in terms of the impact of a one standard deviation increase in the value of the variable. Detailed explanation is offered in Figure 4.2.

A one standard deviation increase in the partisanship of the district (9.3 percentage points), in favor of the incumbent’s party, is associated with an increase of 6.2 percentage points in the incumbent’s share of the final two-party vote.

A one standard deviation increase in the variable measuring challenger spending, however, was associated with only a 3.1 percentage point reduction in the incumbent’s final share of the district vote. Relative to partisanship, challenger spending had a small impact on the outcome of the election.

Generally speaking, each dollar a campaign spends will be less effective than those spent before it. Consequently, like most other statistical analyses of the impact of campaign spending, our regression model uses the logarithm (base 10) of the value of candidates’ spending, rather than the raw amount, to account for the diminishing nature of returns on campaign expenditures. [8]

We estimate that the limited impact of challenger spending diminishes exponentially as their spending increases. For example, the mean level of challenger spending, $446,000, was associated with an improvement in vote share of about 3.8 percentage points over the vote share of a hypothetical challenger spending just $40,000. However, the model suggests that increasing this advantage by just one more percentage point would require spending almost twice as much ($850,000).

More Spending Yields More Votes in Competitive Races

The effects of challenger spending are more apparent in competitive races, but are also slightly lower, even as average challenger spending in these races is higher, further suggesting diminishing returns. Excluding districts with a partisanship of less than 44% or greater than 56% (safe districts) from out model produces different but consistent results. In competitive races, the relative magnitudes of the impact of the significant variables were much closer to one another, and challenger spending becomes the most significant variable. However, incumbent spending remained statistically insignificant to incumbent vote share.

As shown in Figure 4.3, a one standard deviation increase in our challenger spending variable retained comparable magnitude to the full model, while the predictive power and magnitude of most other variables significantly declined. Still, the magnitude of the effects on vote share associated with challenger spending remained low, and the continued relevance of partisanship as well as the increased significance of 2014 [performance over average candidate, or POAC] suggests that district fundamentals and candidate performance are still crucial in close races, despite the high significance of challenger spending.

 Although returns on challenger spending in competitive races also appear to be subject to diminishing returns, the amounts of money spent in many of these races are large enough to be associated with meaningful shifts in the final two-party vote. The data suggest that the mean level of challenger spending in competitive races, $796,000, was associated with a 4.7% reduction in the incumbent’s final share of the vote relative to challengers spending $40,000.

However, due to diminishing returns, we estimate that the highest level of challenger spending observed in the sample of competitive races (over $4 million) generated only a 2.5 % increase in the challenger’s vote share over a candidate spending at the mean ($796,000).

Campaign Spending in Context

Our analysis shows that incumbent spending is not closely tied to the outcome of U.S. House elections when contested by an incumbent. Only challenger spending appears to be related to 2016 U.S. House election outcomes. This means that the only races in which our analysis clearly suggests that campaign spending may have affected the outcome are the eight elections in which a challenger defeated an incumbent.[10] However, our model indicates that the high spending of these victorious challengers ($2.8 million on average) resulted in only a 1.2% increase in vote share over the spending of the average challenger in a competitive race($796,000). The vote share associated with each winning challenger’s spending also suggests that most would have won even if their expenditures were reduced to average levels.

While it is unfortunate that money would play a decisive role in any election over the ideological preferences of voters or the quality of candidates, the data suggest that campaign spending is not as influential in congressional elections as is commonly assumed. The eight races in which we estimate that campaign spending may have altered the final outcome in 2016 represent just 2% of the year’s 394 incumbent-contested U.S. House elections.

In these races, it appears that district partisanship, the national partisan swing of the election, and the incumbent’s recent electoral performance were the most immediate predictors of election outcomes. While it is true that winners of U.S. House elections nearly always outraise and outspend their opponents, our findings reinforce the view that this advantage is a result of donors making investments in candidates they see as likely winners, rather than a causal factor that explains their victories.

Still, methodological difficulties inherent in the measurement of the effects of campaign spending mean that these findings come with qualifications. For example, it may be that incumbent spending appears to be insignificant only because incumbents spend more when they expect a difficult election, reversing the link between spending levels and electoral outcomes.[11] Our model attempts to control for this, but the control may be insufficient. Another factor, which we do not account for, is challenger quality. It may be that politically experienced challengers are both able to raise and spend more money, and achieve better electoral results creating a link between spending and results in our model that is in fact explained by this omitted factor.

Despite these difficulties, the magnitude of the relationship between campaign spending and electoral outcomes we observe is comparable to what has been found in previous studies. The key point, then, for those concerned about the state of American elections or exploring avenues for reform, is that campaign spending is not the primary driver of election results that it is sometimes made out to be. While money may have a wide range of pernicious effects on our political system, it is not to blame for uncompetitive elections and unrepresentative outcomes — its electoral impact only comes at the margins. Reformers looking to tackle these issues should focus on their most immediate cause — the inherent shortcomings of winner-take-all elections, exposed by an increasingly partisan political environment.