Yesterday, Fox News published a breathless exclusive report, using data from something called the Take Back Action Fund, claiming that nearly half of donors using the popular Democratic fundraising platform, ActBlue, are "unemployed" and therefore "untraceable". They also claim that the among Republican donors using the WinRed platform, this number is just 4%. This latter number is just wrong. It doesn't come close to matching the actual data submitted by WinRed to the Federal Election Commission (FEC).
President Trump's press secretary, Kayleigh McEnany amplified the report, along with Senator Lindsey Graham, and other conservative commentators. The President himself weighed in on Twitter:
Fake Liberal Democrat Donors. Money laundering anyone? https://t.co/BAVgQXUS5n
— Donald J. Trump (@realDonaldTrump) September 12, 2020
The implication is that the money that poured into Democratic primary campaigns in 2019 is suspicious "dark money". It's not.
Using the data from the FEC (code available here) for 2019 donations, I estimate the employment rate of ActBlue donors (53.3%) exceeds that of WinRed donors (43.7) by 9.6 percentage points. This shouldn't be surprising. The national employment-population ratio in 2019 was around 61%, meaning 39% of working-age Americans had no job. Combined with the fact that political donors tend to be older and more likely to be retired, around half of donors having no job seems reasonable. That Republicans tend to be older than Democrats also helps explain why a smaller percent of Republican donors have jobs.
Quick note: The Fox News report conflates not being employed with the concept of "unemployment" that is used in calculating the official unemployment rate. The unemployment rate measures the percent of the labor force that does not have a job but is actively looking for a job. The labor force excludes entirely the 39% of the population that does not have a job and does not want a job.
So how does Take Back Action Fund come up with such a small percentage of WinRed donors not having jobs? It's hard to tell. They seem to have only published a few screenshots of spreadsheets, and the numbers on them don't add up to the FEC's totals. But it seems the biggest reason is that they excluded retired people (40% of WinRed donors) from the WinRed data. It is not possible to do the same with the ActBlue data, where only 2-3% of donors can be identified as retired. There is a reason for this: ActBlue and WinRed are conduits for political donations, which means that under campaign finance law they must collect certain information from everyone who donates via their platform. The required information includes each donor's occupation and employer, which is not possible for people who have no job. ActBlue's data looks different from WinRed's because they approach this problem in different ways. ActBlue's form for collecting this data looks like this:
If the donor indicated that they were not employed, ActBlue fills in "NOT EMPLOYED" on the massive file it sends to the Federal Election Commission (FEC). If they select "yes", they are required to enter something in the occupation and employer fields. Presumably, most donors who are retired would just select "no" and be counted as "NOT EMPLOYED". The following chart shows the percent of people selecting this option by month, as well as those who indicated that they were employed then entered something in the occupation field suggesting that they were not actually employed.
WinRed, which only started collecting substantial numbers of donations in the second half of 2019, used a similar design, except it explicitly mentioned retirement:
And in October, the form changed to remove the "unemployed" wording on the button, leaving it only as "I'm retired".
It's hard to know how people without jobs who are not retired would navigate this form, because unless you select "I'm retired", the occupation and employer fields are required. Following the change, the number of people listed as "RETIRED" in the WinRed data spikes, while the number listed as "NOT WORKING" takes a steep dive. This suggests that some donors without jobs are likely selecting the "I'm retired" option to escape having to enter an occupation, whereas the rest are probably entering something into the occupation field to indicate they don't have jobs.
The upshot of this quirk in the donation forms is that the most retirees and probably even some non-retirees are classified as "RETIRED" in the WinRed data. In the ActBlue data, retirees are almost entirely counted as "NOT EMPLOYED" unless they take the odd step of indicating that they are employed, then entering "retired" in the occupation or employer field.
It's important to understand that this data is easy to access on the FEC website, and that it is not hard at all to calculate these comparable estimates of donor employment rates. Anyone with the means to download and analyze the nearly 23 GB of data also knows how to create some descriptive tables and figure out why the two organizations filings look so different from each other. I don't think the people at Take Back Action Fund are so bad at data analysis that they missed this. I think they created a report with the intent to deceive Fox News readers. And the President and Republican politicians are more than happy to play along.
Note: I've made my code and data available on github.
