Boston area residents voting in the September 4th primary had very little indication of who the front-runners in this election cycle might be. Only the race for the Massachusetts 7th Congressional District had been recently polled – a WBUR/MassINC analysis in early August showed Capuano in the lead at 48%, with Pressley trailing at 35%. Other critical races, such as Suffolk County District Attorney or the 15th Suffolk District, had no polls at all to indicate the temperament of voters. With very little official polling available, much of the speculation around supposed front-runners was based on informal straw polls and assessments of campaign war chests.
Given the lack of polls in this primary election and the inaccuracy of the polls that were available, we sought to answer this question:
Does the concentration of campaign contributions within the voting area predict the outcome of the election?
15th Suffolk House of Representatives
We first analyzed donations to Representative Jeffrey Sánchez and challenger Nika Elugardo to evaluate the concentration of contributing households within the voting district, and we were shocked by the findings. Although Sánchez had raised more than three times as much as his challenger, Elugardo appeared to have a much higher saturation of contributions within the area where eligible 15th Suffolk voters lived.
This finding was significant because contributors to political campaigns can be considered to be a level above super-voters. While their vote is almost certainly assured, one can reasonably assume this level of activated voter is also having conversations with their friends and neighbors about why they’re passionate about their candidate. As we now know, Elugardo went on to defeat the incumbent by 52% on September 4th, spending only $22 per vote, while Sánchez spent $67 per vote.
Suffolk County District Attorney
Next we examined the Democrats in the Suffolk County District Attorney race. While this data set was more complex due to the number of candidates, we found a similar result. Rachael Rollins had a higher concentration of contributors throughout Suffolk County and the distribution pattern was more homogenous.
Rollins won the Democratic primary for Suffolk County District Attorney by earning 39% of the votes. As my collaborator Bobby Constantino reported, Rollins also benefited from the financial support of the neighborhoods most disproportionately impacted by incarcerations and detentions – even though these areas have lower median incomes than other areas of the county. Rollins spent roughly $6 per vote, while Henning spent $24, Carvalho spent $10, McAuliffe spent $28, and Champion spent $7. While white candidates may have access to larger bases of financial support than candidates of color, it does not necessarily ensure them electoral victory.
By plotting bimonthly deposits into campaign accounts, it was clear that Rollins was surpassing her opponents in the weeks preceding the election. It’s also worth noting that the last deposits before the election mirrored the election results precisely – with Rollins in the lead, trailed by Henning, Carvalho, McAuliffe, and Champion.
Massachusetts 7th Congressional District
Next we applied our theory that campaign distribution patterns predict outcomes to the Massachusetts 7th Congressional District. Even though the incumbent Michael Capuano raised twice as much as Ayanna Pressley, when we mapped contributions within the parameters of the district, it appeared that Pressley had a higher concentration of support throughout the district, while Capuano’s support was concentrated in the north.
Pressley would go on to defeat Capuano by receiving 59% of votes, proving that the most recent WBUR/MassINC poll of the race was wildly inaccurate at predicting the outcome of this election.
Another stunning finding consistent with the other victorious women of color mentioned above: Pressley accomplished her win by spending only $13 per vote, while Capuano spent $41 per vote. With targeted strategies and strong grassroots support, candidates are winning elections despite being massively outspent by their challengers.
9th Suffolk House of Representatives
After the primary election, we sought to validate our theory by applying it to another local election. In Suffolk’s 9th District, Representative Byron Rushing was challenged by Jon Santiago and Suezanne Bruce. While much less cash was spent in this race, the pattern of contributions also revealed a higher number of donations within the voting area for newcomer and winner Jon Santiago. Incumbent Byron Rushing spent $19 per vote, Suzeanne Bruce spent $3 per vote, and Jon Santiago spent $10 per vote – again showing that newcomers can topple incumbents without outspending them.
This sample set is small and more rigorous statistical analysis of past and future elections needs to be performed to confirm our theory; however these findings strongly suggest that outcomes may be reasonably predicted by examining the concentrations of campaign donations within a district's voting area.
Further analysis also needs to be performed to determine how much the increased interest in the 2018 election cycle is influencing both campaign contribution patterns and constituent turnout. In the meantime, journalists, pundits, pollsters, and candidates would be wise to incorporate readily available campaign finance data to evaluate voter support and predict electoral outcomes.
We exported campaign contributions from the Massachusetts Office of Campaign and Political Finance for state elections or the Federal Election Commission of the United States of America into Microsoft Excel Spreadsheets. Data was then imported into Google Maps. Using Adobe Illustrator, we traced the official district map and placed it over the corresponding area on the maps of contributions. We deliberately did not examine or include the amount of each donation – we only used the address and zip code data (P.O. Boxes were changed to the address of the closest corresponding post office). To calculate dollars per vote, we divided total campaign expenditures by the total number of votes cast, correcting for known reporting errors in the OCPF database. This project was done in collaboration with Bobby Constantino. Special thanks to Krista Magnuson for her copyediting expertise.