Unofficial 2016 election returns + voter turnout + campaign finance info

I, for one, welcome our new Alt-Right Grimace overlords.

I, for one, welcome our new Alt-Right Grimace overlords.

Hi everyone,

The torturous 2016 elections are two weeks behind us, and all but two federal Congressional races have been decided.

Because of one of the quainter provisions of Article I (stipulating that states regulate elections and compile the returns themselves), it will be many more months until we know exactly how many people voted for every candidate in every election, and still many more before the Federal Election Commission compiles these compilations into one usable spreadsheet. In light of this, I’ve decided to make available the unofficial election results for presidential (both national and state-wide), Senate, House, and gubernatorial races.

DOWNLOAD THE DATA (Excel 2007 format). See the appendix below for a codebook describing what’s in the file.

Some notices:

  • The conventional wisdom for about 10 years is that Democrats will enjoy an increasing advantage in presidential elections as the white share of the voting population declines. Indeed, this is borne out by the fact that Democrats have won pluralities of every presidential election since 1992, with the exception of 2004 (an odd duck because of its immediately post-9/11, Iraq War setting; had the election been held even a few months later, Bush would’ve lost). There is no reason to doubt this, but note the anti-majoritarian structure of the Electoral College and Congress (where minorities are packed into relatively small numbers of densely-populated districts, allowing elections like 2012 where Democrats win a majority of House votes but a small minority of seats) negates this advantage almost completely. The conscious efforts of high-profile Democrats like Chuck Schumer to appeal to high-income professionals at the expense of traditional Democratic constituencies backfired, since the latter are disproportionately concentrated in swing states and the former are disproportionately concentrated in the uncompetitive coasts. Hillary Clinton thus won 1.5 million more votes than Trump but lost states that have been solidly Democratic since 1988.
  • Almost everything that pundits said about this election and previous elections is false. They said that Trump’s massive deficit in fundraising (in the summer, his campaign had only $1 million on hand–enough to maybe win a House seat!) and horrible campaign messaging would bury him. They failed to remember the 2012 election, when Karl Rove’s PAC spent perhaps half a billion dollars of rich people’s money to retake Congress and the presidency, to absolutely zero effect. The pundits also missed another insight that sociologist Paul Lazarsfeld shared back in 1948: media coverage of campaigns doesn’t matter. Perhaps to their credit, most voters do not pay attention to the political news cycle or watch Mark Halperin bumble about “who won the week.” Hillary lost because she though that Trump’s gaffes would do all the work for her; her campaign foolishly made huge ad buys to raise the message that Gulf War era Republican geriatrics supported her, and that Trump has a virulent case of verbal diarrhea, all at the expense of get-out-the-vote drives. In so doing, she primed voters to think in terms of Trump’s sources of appeal: his scorn for the mainstream Republican party and the Beltway, and de-emphasis of traditional GOP tropes like Constitutional fetishism. The feared “ground game” on the part of the Clinton campaign was ultimately underfunded, mismanaged, and in some cases actually turned out Trump supporters to the polls.
  • Trump outperformed Mitt Romney in every demographic (including minorities, by a substantial margin) except college-educated whites. Hence, “white working class” has replaced “Reagan Democrats” (1980; 1984), “NASCAR dads” (2004) and “soccer moms” (1996; 2000) as the nauseating shibboleth of punditspeak in 2016. However, the focus on the “white working class” ignores that A) Trump’s appeal over Hillary was general, and not exclusive to supposed Archie Bunker types; B) over 40% of Trump voters are college-educated (to make a historical parallel: Christopher Lasch noted that Reagan’s core constituency was not hard-hat reactionaries as the media likes to portray, but rather suburban professionals); and C) the majority of the working class of every ethnicity did not vote. Trump won about as many votes as Romney did. The takeaway is that the Clinton campaign simply did not offer compelling reasons for members of Obama’s electoral coalition to show up to vote. A message of “I’m competent, and competent people in think tanks like me” is hardly inspiring to young people, and was calculated to fail when (say) Clinton so ineptly handled the email server scandal.
  • The Democratic party is in trouble. Its share of Congressional seats, state executives, and state legislatures is as low as it was in the 1920s. Perhaps worse, the party lacks any young, up-and-coming stars with the national appeal necessary to win elections. The Democratic National Committee has been wholly discredited except as a source of jobs for newly-minted JDs with family connections. An incalculable number of historical tendencies (the destruction of the union voting bloc; the miserable performance of the economy that created divisions in the New Deal coalition since the early 1970s; the erosion of distributional politics [e.g. earmarks] that guaranteed 40 years of Democratic majorities in Congress) brought us to this point, such that no individual or group of individuals is responsible for the party’s predicament. The Clintons still deserve blame for alienating voters and ineptly engineering the Republican realignments of 1994 and 2016, and as memories of the modest economic bubble of the 1990s recede so too will their historical standing.
  • Polls are becoming increasingly untenable as ways to measure public opinion owing to abysmal response rates. This made it possible for Hillary Clinton to lose Wisconsin when not a single poll at any point in the election cycle showed that Trump was in the lead there. In the coming years, we’ll likely see their replacement with Internet data mining.
  • Bernie Sanders would almost certainly have won. Valence (candidate pluses and minuses), not demography, decide the outcome, and where Clinton has for 25 years been one of the most hated politicians in America, Sanders overwhelming polls as likable. That being said, it’s not the case that he in any way cost the election for Hillary, as many journalistic postmortems have held. (He somehow “primed” voters to dislike Hillary, pundits claimed, when his no campaign chivalrously avoided mentioning the baggage that proved fatal in so many parts of the country: the emails, the Clinton Foundation’s influence trading, the general perception of corruption, etc.). Outside of Twitter (which only few people use regularly) and people with media jobs, no one was talking about Sanders in the context of the election after May. The only certainty in American politics is that what op ed writers think is important is irrelevant.

Appendix: Codebook

Variable Name Variable Description
Election ID Unique identifier for the race. Jurisdiction + office + election type
Election Type G = regularly-scheduled election for a term of office that has expired; S = special election for unexpired term
Election Date Date of election
State Name Name of the state (or, if national presidential returns, country)
State Postal Code State postal code
State ICPSR Code Two-digit Interuniversity Consortium for Political and Social Research (ICPSR) code for the state. Left digit = region (0 = New England, 1 = Mid-Atlantic; etc.)
Congressional District # Congressional district number. At large districts (e.g. Alaska, Vermont, Wyoming, etc.) are coded “1”
Jurisdiction Population Population of the jurisdiction per the most recent American Community Survey (ACS) in my files
Jurisdiction Voting-Age Population Voting age (18+) population (VAP) of the jurisdiction per the most recent ACS in my files
Office P = President; H = House; S = Senate; G = gubernatorial. In 2016 all Senate races were for Class 3 seats
Open Seat O = no incumbent running in the election
Voter Turnout Non-blank, unspoiled ballots cast for a candidate in the election
VAP Rate Number of ballots cast divided by VAP
Effective Numer of Candidates Commonly-used measure of election competitiveness.  ENC = Σ(ci2)-1, where c is the vote share of the votes of ith candidate
Party Party of the candidate as it appeared on the ballot
Write-In W = write-in candidate
Incumbent I = incumbent
Place Place in which the candidate finished (1 = popular vote winner; 2 = runner-up; etc.)
Vickrey Candidate ID Candidate ID in my database of candidates going back to the 1960s. Has no external significance
FEC Candidate ID 9-digit Federal Election Commission ID for the candidate in the 2016 election cycle. Omitted for most write-ins, e.g. Bernie Sanders
FEC Principal Campaign Committee ID 9-digit FEC campaign committee ID. Useful for merging in filing-level campaign finance information
Candidate ICPSR Code Candidate ICPSR code, where applicable. Taken from Keith Poole’s files on VOTEVIEW. Useful for merging in DW-NOMINATE scores
Voted Received (Fused) Votes received by the candidate for all parties, for instance in states that allow fusion voting (commonly practiced in CT, NY, and SC)
Vote Share (Fused) Fused votes divided by turnout
Votes Received (By Ballot Party) Votes received by the candidate by party (e.g., how many New Yorkers voted for Hillary Clinton running as a Working Families candidate)
Vote Share (By Ballot Party) Ballot party votes divided by turnout
Campaign Receipts Money raised by principal campaign committee up to 10/19/2016
Campaign Disbursements Money spent by principal campaign committee up to 10/19/2016

About cvickrey

Clifford Vickrey spends his days confounding the wise.
This entry was posted in political science, politics, Uncategorized. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *