Late at night, a good Samaritan happens upon an intoxicated man crawling under a streetlight, studying the ground. The drunkard says that he is looking for his keys, and our helpful bystander joins the search. After a few minutes, the helper asks, “Are you sure this is where you dropped them?” “No,” says the drunk, “I probably dropped them by the bar down the road. But the light’s much better here.”
The phenomenon of the drunkard’s search, often attributed to Abraham Kaplan, afflicts much empirical inquiry. Researchers seeking quantitative rigor too often head for the available data, fire up STATA, and only as an afterthought offer a few generic caveats about the limitations of the data for the topic at hand.
Campaign finance analysis is not immune from this problem. In reading some recent studies over the weekend, I noticed a few recurring lapses.
Consider the evaluation of electoral spending by entities other than candidates and parties: “outside” spending, for short. Frequently, studies look to data from the Federal Election Commission (and equivalent state sources) to track the rise of independent expenditures since blockbuster Supreme Court cases like Citizens United, which allowed corporations (and, by logical extension, other non-affiliated groups) to engage in such activity without limits.
There is no question that independent expenditures have increased dramatically. But most people do not realize -- and most analysts do not emphasize -- that the term “independent expenditure” has a technical legal definition. When the FEC reports data, it considers an “independent expenditure” to be a communication expressly advocating for the election or defeat of a clearly identified federal candidate, with phrases like “vote for” or “elect.”
If, in gauging changes over time, the only spending of interest is a communication expressly advocating with respect to a candidate, then limiting a data search to independent expenditures is perfectly appropriate. But, if the research question concerns electoral spending more generally, the data on independent expenditures represents a mighty narrow streetlight.
More sophisticated analysts of electoral spending will add, as this study does, “electioneering communications.” These communications are also legally defined: they cover certain types of broadcast ads referring to clearly identified federal candidates within 60 days of a general election or 30 days of a primary. An election-eve TV ad explaining that Congressman Levitt hates puppies, and exhorting viewers to call Congressman Levitt to tell him that he’s an awful person, would be an electioneering communication. And though it might not show up on FEC reports of “independent expenditures,” a researcher examining trends and shifts in electoral spending would probably want to take such an ad into account.
And yet, the electioneering communications streetlight is only slightly broader. Legal requirements for reporting data with respect to these ads only arose in 2002. When you see a jump in the amount of electioneering communications in 2004, that does not indicate a sudden appearance of brand-new Levitt-hates-puppies ads. It meant that, for the first time, such ads fell under the streetlight. How much spending was there on ads in 1998 or 2000 or 2002 that would now meet the definition of “electioneering communications”? FEC data leaves no idea.
Indeed, this is only the start of the limits to our understanding. Right now, there is much debate about so-called “dark money:” electoral spending that is disclosed to the FEC, with donors who aren't. But, there is also a pot of “darker-than-dark” money: electoral spending that the FEC doesn't really know about at all. If an entity does not have the election or defeat of candidates as its primary purpose, spending that is neither expressing advocacy nor on the air as an election approaches is entirely outside of the current electoral disclosure regime. (Some is reported to the IRS or other government bodies, but it’s not compiled in a way that permits useful apples-to-apples comparison.)
Labor spends millions on pre-election canvassers to discuss candidates’ records on union pensions? A pro-choice group spends millions on pre-election mailers touting candidates’ support for abortion? A conglomerate puts up a TV ad 31 days before the primary railing against a candidate’s sexual peccadilloes? For the FEC, this is terra incognita. Here be dragons.
To be clear, I am not advocating for full disclosure of all of this activity. I am, however, advocating for more frequent acknowledgment of these known unknowns, particularly in empirical analysis of the changes that changes in the legal regime have wrought.
We don’t know how much outside campaign spending there is. And we don’t know how much there has been. We know only how much is, or was, under the streetlight. Usually, that has meant independent expenditures and electioneering communications. There is other electoral spending out there.
The corollary point is that we’re not sure how much substitution there has been. It may be that there were vast quantities of activity outside of the streetlight that have since come under the streetlight: money may have shifted from untracked activity to tracked activity as tracking has expanded and legal barriers have fallen. It may be that the spending we are now seeing is really new. But, it would behoove analysts to recognize the limits of their ability to tell one from the other.