Reporting recipe: How to investigate out-of-state influence in 2018 races

Introduction

Control of state governments — and with it influence in the legislative redistricting following the 2020 census — is up for grabs in dozens of executive mansions and state legislative chambers across the country in this fall’s elections.

Already people, PACs, corporations and unions have contributed more than $173 million in out-of-state money to gubernatorial, state house and state senate candidates.And that’s just money sent directly to candidates. It does not include hundreds of millions more spent by outside groups on things like television advertising to support or oppose candidates.

To better understand the extent of out-of-state giving — where it’s going, how it compares with previous cycles, and which party has an advantage — Center for Public Integrity reporters Rui Kaneya and Joe Yerardi did extensive data analysis and on-the-ground reporting.

What can local journalists do with this data?

The code and data behind this story’s analysis, along with a data dictionary is publicly available to download. It is our hope that journalists located in states with particularly interesting giving patterns will use these resources as a starting point for their own reporting on the flow of out-of-state money in elections.

Some potential questions this data can help answer:

• Have candidates in my state seen a surge in out-of-state giving as compared with prior cycles?

   

• Are one party’s candidates pulling in much more out-of-state cash than the others?

   

• Is a particular candidate or handful of candidates especially reliant on out-of-state donors?

   

Important things to keep in mind when using this data

This data comes from the National Institute on Money in Politics. NIMP collects campaign finance data from all 50 states, adds additional information such as the industry in which a donor works and provides it as downloadable data. As valuable as this service is, there nevertheless remained several caveats we had to keep in mind as we analyzed it.

First and foremost, these campaign finance reports may not be the most up to date. Candidates must file their campaign finance reports only periodically and all states have different deadlines, so the latest contributions may not have been reported yet. In states with particularly arcane campaign finance systems (think paper filings only), NIMP researchers manually enter information from the reports into their databases. This time-consuming process means that these states tend to be the most out of date. NIMP prioritizes states with the most competitive, high-profile races (think governors races) over states with lower-profile contests.

The contribution numbers we are publishing reflect the data in NIMP’s system as of Oct. 1, 2018.

One of our primary questions was whether out-of-state contributions have increased as compared with prior cycles. To ensure that we could make accurate comparisons across cycles, we implemented four filters.

First, we filtered out unitemized contributions. These are contributions from donors who have given so little that campaigns need not report details on who’s giving. They’re reported as a single total given by all the donors on each campaign finance report. Due to the lack of detail, it’s impossible to tell what proportion of each reporting period’s unitemized contributions originated from inside or outside the candidate’s state.

Second, we filtered out prior cycle contributions made after the latest month for which each state had data this cycle. For example, Florida data was only available through August. So we filtered out Florida contributions for the 2014 and 2010 cycle that occurred after August.

Concerned that the most out-of-date states would only reflect a small portion of those states’ eventual contributions pool and thus skew our results, we implemented a third filter to eliminate all contributions from the seven states where the latest contributions shown in NIMP’s system occurred in 2017. As 45 states are holding gubernatorial or state legislative elections this cycle, this left us with data from 38 states.

Similarly concerned about skewing our results with data from also-ran candidates raising a very small amount of money, we filtered out candidates who raised less than $1,000 in their race.

With these filters in place, we found that out-of-state giving was up more than 14 percent as compared with this same point in the 2014 cycle and was up a whopping 50 percent as compared with 2010.

Hanging over this year’s elections is the role that some state offices play in the redistricting that will occur after the 2020 Census. We were curious whether out-of-state donors were particularly interested in governors and legislatures with influence in that process. Despite our hypothesis that they would be, we found there was not much of a difference between the proportion of out-of-state money going to the one set of candidates over another.

Lasty, a quick explanation on what we mean by “in-vs.-out-of-state.” NIMP determines whether a contribution is in-state or out-of-state based on the location of the candidate receiving the contribution and the reported address of the contributor.

What data are you publishing?

We are making data and a detailed data dictionary available in this project’s GitHub repository.

What follows is a brief overview of the files being made available.

The “National” file contains data on national trends in out-of-state giving.

The “Candidates” files contain data on out-of-state giving trends to individual candidates.

The “States” files contain out-of-state giving trends state-by-state.

The “Contributors” files contain contributor-level out-of-state giving trends.

Journalists can use this data as a springboard for further reporting, using NIMP’s search tools and/or the campaign finance reporting system for their state.