The explosive growth in the cost of U.S. elections has continued to raise questions about the role of money in politics and its degree of influence on elections. Recent election cycles have shattered campaign fundraising and spending records: the unprecedented 2020 U.S. election cycle cost nearly $14 billion in presidential and congressional campaign spending, exceeding the spending amounts witnessed in the 2012 and 2016 U.S. federal campaign cycles combined. For context, the amount of spending poured into 2020 U.S. races alone would eclipse the GDPs of nearly one in three countries globally.
Over the past several decades, significant changes to campaign finance laws have led to dramatic shifts in the scale of money in elections. In particular, the Citizens United v. Federal Elections Commission Supreme Court ruling in 2010 paved the way for unrestricted political spending from corporations and other large donors. Opponents of large sums of money in politics have clamored for years about the need for campaign finance reform, calls that have intensified in the wake of the Citizens United outcome. They argue that money fundamentally distorts elections to favor special interests, including those of the wealthy elite, thereby inherently threatening the sanctity of America’s democracy.
But to what extent does money actually play a significant role in influencing election outcomes? Countless hypotheses have been posed on this topic, including some who argue that money has direct influence in distorting election outcomes to favor the wealthy and others who assert that money and election outcomes may display a correlative, but not necessarily causal, relationship. To explore these explanations further, I sought to answer the question empirically. In this first of a multi-piece series on analyzing campaign finance, I examine the relationship between political spending and the election outcomes for the 2020 U.S. Senate races. I leveraged a publicly available dataset containing campaign finance data for 2020 U.S. Senate races from OpenSecrets through the Center for Responsive Politics.
The table below displays the 35 U.S. Senate races in the 2020 election cycle, including a special election in Georgia, along with the amount raised, amount spent, and vote share for the state’s winning candidate and losing candidate. (Note: A star by a candidate’s name denotes incumbency.) It also includes columns describing the differential between the two candidates based on the three parameters previously mentioned. Based on this data, I attempt to address the following questions: What proportion of U.S. Senate races were won by the candidate that outraised or outspent their opponent? Given that some races are substantially more expensive than others, are there any noteworthy qualitative factors influencing the cost of a race and the likelihood of a particular election outcome? What is the relationship between fundraising margin and vote margin?
Figure 1. Campaign Finance Data for 2020 U.S. Senate Elections
Initial findings from the empirical analysis suggest that there is some evidence to support the notion that money may influence electoral outcomes. In 25 of the 35 (71%) U.S. Senate races during the 2020 election cycle, the candidate that outraised and outspent their opponent won the election; if money had absolutely no influence on election results, one could reasonably expect a proportion closer to a 50-50 split.
Next, I analyze the relationship between the amount that a candidate spent on their election and the vote share they received during the election. As demonstrated in Figure 2, which applies a log transformation to the x axis, a simple linear regression model yields a weak positive correlation between these two parameters, suggesting that campaign spending may have a small, but non-trivial, influence on a candidate’s vote share. However, this analysis is itself limited, as there are many underlying factors that influence the vote share a candidate receives, notably the partisan breakdown of a given state. Indeed, as seen in Figure 3 below, a clear, intuitive trend emerges: the most expensive races are also likely to be the most competitive races.
These results already suggest a more complicated relationship between money and its influence on electoral outcomes. Notably, incumbency and polarization may be an even greater driving force than money in influencing election results. For example, studies have long suggested that incumbents benefit from structural advantages that make them disproportionately likely to win re-election. While this trend can be attributed to multiple factors, money plays an important role: incumbents have a head start in amassing a war chest to outraise and outspend their candidate, while also benefiting from greater name recognition for prospective donors. Indeed, as OpenSecrets finds, 26 of the 31 (84%) incumbent U.S. Senate candidates won re-election, while only 5 were defeated by challengers. Of the 5 incumbent candidates that lost re-election, including Democratic Sen. Doug Jones and Republican Sens. Martha McSally, Cory Gardner, David Perdue, and Kelly Loeffler, all but Sen. Doug Jones were outspent by their challenger. However, these five elections occurred in states where the political party of the incumbent Senators that lost their election were the opposite of the party that won the Presidential race in that state: the previously mentioned Republican senators, who represent Arizona, Colorado, and Georgia, all lost re-election in a state that Democratic President Biden won, while Democratic Sen. Doug Jones, who represented Alabama, lost re-election in a state that former Republican President Trump won. These outcomes suggest that the state’s partisan divide may have explained the election outcomes to a greater degree than the amount the candidate raised. While money may have some degree of influence over an election outcome, the relationship is complicated. Further research on the role that several other key factors, including incumbency and name recognition, will be necessary to continually develop a more nuanced understanding of money’s role in politics.
Lastly, I examine the relationship between spend margin (defined as the winning candidate’s amount spent minus the losing candidate’s amount spent) and vote margin (defined as the winning candidate’s vote share minus the losing candidate’s vote share). Figure 4 depicts all 35 U.S. Senate candidates in 2020, displaying no discernible positive or negative relationship between spend margin and vote margin. While many candidates who significantly outspend their opponent by tens of millions of dollars may be expected to beat their opponent by significant amounts, there was little evidence to suggest a broader trend. In fact, some of the races with the largest differentials in amount spent between the two candidates were also some of the most competitive. This trend cannot be readily explained by a simple narrative, such as one implying that more money necessarily means more unfair elections.
This piece is intended to be the first of a multi-part series analyzing campaign finance data and the role of money’s influence in politics. Exploring the trends highlighted in this investigation of 2020 U.S. Senate elections across other races, including U.S. House and state legislative elections, and previous election cycles will help further refine this analysis moving forward.
While this initial analysis has raised more questions than it has answered, one thing remains clear: complex dynamics underlie the relationship between money and elections, and simple, black-and-white narratives do not necessarily exist. While campaign finance reform is an important step to curbing money’s role in elections, it is also important that advocates continue to investigate other underlying factors that may influence elections in favor of various special interests.
Stay tuned for further analysis by Charles Hua on the impact of money in politics.
Image by Sharon McCutcheon is licensed under the Unsplash License.
