My day job is with a data analytics software company, and sometimes that job allows me to play with data and make pictures. Today I spent looking at the full FEC contributor data set from fec.gov. For some reason, the available data only goes to February 1st, so if anyone know where I can find more recent data, please let me know. Size is no object (I work with big data all the time) I just need to figure out where it is.
I made a lot of pictures, but I managed to fit the most interesting ones into a single “dashboard”. It was originally in a vector format (pdf) but Kos doesn’t seem to accept that, so I’ve had to rasterise it (which is why it is a bit blurry). But here it is:
Contribution Overview
The top visualisation shows the number of contributors, the average contribution size and the total amount of money raised for all the candidates who are still in the race. Some interesting findings:
- Clinton has almost as many contributors as Sanders, and both of them together blow away the rest of the field (GOP and Green). The fact that they have similar numbers will be important later.
- The average contribution size of the Clinton camp is about 4 times that of the Sanders camp. This suggests that Clinton appeals to more affluent Democrats.
- Surprisingly, though, Clinton does not have the highest average donation among the remaining candidates: That honour goes to Kasich of all people. This suggests that the two “establishment” candidates have the backing of the “heavy hitters”.
- Donald Trump has almost no contributors, but those who do give to him seem to give a fair bit. This supports the argument that he has a wider base of support than angry, unemployed white men (small business owners?)
- The Green Party candidate (Julia Stein) has few contributors, but the average is still higher than Bernie’s, suggesting that her appeal is also to the relatively well-off. This is not surprising — only those with a lot of security can afford to fund third-party candidates.
- Sanders has the smallest average donation size, but the roughly $325 average per-contributor is much larger than the much-quoted $27 figure. Even if you break it down into average transaction size (not shown), it still comes to $75. But even this is still much smaller than Clinton’s average of $470.
Fundraising Over Time
I really wish I had more recent data because the most interesting line on the time trend is Sander’s fundraising acceleration after the Iowa caucuses. Everyone else is showing flat growth, although Cruz seems to be picking up a bit. But that corner on Bernie’s graph is impressive.
The other thing to note is the massive lead that Clinton has over everyone. If you ever wondered how much truth there is to money buying elections, this will be a good year to find out. For those wondering about Bush, he only hit about $30M on the graph — well below Clinton, and even below Cruz by the end of January.
Bernie’s Endgame
The last map is the most speculative. To make this, I computed the number of Sanders supporters in each state that has yet to run a primary or caucus and divided it by the total number of supporters of any other Democratic candidate (basically Clinton — the other four are noise.) (Did you forget about Lawrence Lessig? I thought you might have!) I then expressed it as a percentage and plotted it as a blue-red diverging colour encoding. This percentage is a projection of Sander’s expected vote totals by state. This is a reasonable thing to do because they both have similar numbers of contributors.
(Brief visual nerdiness: Red and Blue are well separated in human visual perception, so I used them to make sure the contrasts are visible close to the middle. Red/Green works too, but I was worried about colour blindness. Making Bernie blue was because they both begin with “B” (and Hillary has more “R”s) but also because I am a Bernie supporter and it amused me to use the GOP colour for Hillary...)
This is why the Sanders campaign has been claiming that the remaining races favour him. Looking at New York, I’m not convinced, and California is too close for comfort (52%), but the claim is in line with my rough model.
Now that model isn’t great. I reasoned originally that anyone who puts money on the line is a likely voter, but It doesn’t do well in close states like Ohio. Still it gets the broad picture right, and shows why the Sanders campaign is making this claim.
Bottom Line
So here is the situation oh my fellow Sandernistas. We are being outspent by a factor of four, including a lot of people who can give more than most of us. We are a democratic campaign that eschews donation inequality, so we need to make up for it in volume. Most of you have already given what you can, but I’d ask anyone who hasn’t to please chip in.
Yet in spite of being outspent, we are also showing that money isn’t everything. Our “widow’s mites” are taking us close to parity with the most well funded campaign in the country and making them sweat. Let’s keep them sweating by getting out the vote and refusing to give in to all that money.
Thursday, Mar 17, 2016 · 10:43:36 PM +00:00
·
Hawkfish
Some of the comments are pointing out that the data does not contain donations under $200 and I suspect they are correct. When the numbers did not come out as expected, I did check to see if there were any small donations and there are — in fact the distribution is pretty Poisson-like, which is what I would expect.
The other links provided are summaries and I was looking for the raw data. So if anyone knows where to find it, I would appreciate it. I don’t know if I will be shooting my mouth off any more, but I have a professional interest in investigating ways to clean it.
Lastly, my comments about spending are clearly off, so they should be ignored. I do think some of the other patterns are valid, but I am less confident than I was...