Why New Orleans Won't Be Enough to Get Derrick Edwards Over the Top
There’s some talk that a competitive mayoral race in New Orleans could significantly boost the chances for Democratic candidate Derrick Edwards to win the statewide special election for State Treasurer. Reporting by Julia O’Donoghue of nola.com does a thorough job exploring that argument and its weaknesses. Folks interested in Louisiana politics would do well to read it. Seriously, go read it now. I’ll wait…
...okay, now let’s take a look at some numbers to further unpack why New Orleans won’t turn the Treasurer’s race.
Why do some people think a New Orleans mayoral contest helps the Democrat running for Treasurer?
On its face, the argument makes sense. New Orleans is a deep blue city in a mostly red state. It went for Hillary Clinton with 81 percent of the vote in last year’s presidential contest even as Louisiana went 58 percent for Donald Trump statewide. A competitive mayor’s race fuels higher local turnout, which means more Democratic voters are in the booth and likely to also flip the lever for Derrick Edwards in the Treasurer’s race. This could be especially important in a statewide election that has generally low turnout outside of New Orleans. The Secretary of State predicts statewide turnout as low as 12 percent of registered voters. High turnout in New Orleans from the mayor’s race could allow the city to punch above its weight in the statewide balloting.
Will the mayor’s race boost turnout?
Yes. First, let me own up to a miss I made in projecting turnout in New Orleans during the October primary. I estimated the mayor’s race in New Orleans to boost turnout there by about three percentage points, or roughly 8,000 votes. That was waaaaaaaaay below the likely impact of the mayor’s race.
Turnout in New Orleans was about 32 percent of registered voters. In the Treasurer’s race, 29 percent of registered voters in New Orleans cast ballots – or 74,409 votes. Statewide, only 13.5 percent of registered voters cast ballots in the Treasurer’s election. If New Orleans had the same turnout rate, there would have been only 34,812 votes for Treasurer coming out of the city. So, rather than 8,000 additional votes, there were 39,597 additional votes. [If you care to read more about why my approach didn’t work, I added a brief paragraph at the end of this post.]
In fact, the mayor’s race may have yielded even more additional votes than that because New Orleans has typically had lower turnout than the state during statewide elections over the past decade (figure 1). Across the 13 statewide elections from 2007 through 2016, the share of voters casting ballots in New Orleans is seven percentage points lower than the statewide share on average. That has narrowed in recent years, but even looking since 2012, the gap is about three points. This October it was 18 percentage points higher than the state.
So, yes, the New Orleans election will drive turnout up. If the primary is instructive, there will be at least 74,000 votes cast for Treasurer in New Orleans perhaps higher. There were more early votes for the runoff than for the primary in New Orleans, even as early votes declined statewide.
Will it be enough to turn the Treasurer’s race?
Almost certainly not. Edwards has a steep climb outside the city. He carried only 24 percent of the primary vote outside New Orleans. He would need that to come up along with a rise in New Orleans turnout to get close to a majority.
Both of these are difficult. Indeed, the latter is probably impossible in the sense that there simply are enough votes in New Orleans to make up for a significant vote share gap outside the city. New Orleans has seen a dramatic decline in turnout for mayoral elections over the past 23 years. Figure 2 shows the total votes cast and the turnout rate during this period. Eighty to ninety thousand total votes looks like about all you can expect.
To see how it all works together, consider two optimistic scenarios (optimistic from the Edwards point of view).
Scenario #1: Statewide turnout drops to 12 percent but the number of votes for the Treasurer’s race in New Orleans increases 8 percent to about 80,000 (roughly equivalent to the local increase in early votes from the primary to the runoff). This is optimistic because it probably requires the total votes in the mayoral election to shoot up to the upper 80 thousands. Remember not all voters in the mayoral election will also vote for Treasurer, there was a gap of about 6,000 votes in the primary.
Twelve percent of voters casting ballots in the Treasurer’s race is about 360,000 total votes – if we say 80,000 are from New Orleans, then we have 280,000 from the rest of the state. A candidate would need more than 180,000 to win.
Let’s also assume that Edwards does slightly better in his vote share from New Orleans during the runoff. This is reasonable because the marginal additional voter is more likely to be Democratic. He carried New Orleans with 62 percent of the 74,000 votes cast there in the primary. So, with 80,000 votes cast in New Orleans for State Treasurer, maybe he carries the city with 65 percent of the vote (basically sweeping the additional 6,000 voters at 95 percent), for about 52,000 votes.
Outside New Orleans, if Edwards stays at 24 percent of the vote in the primary, he gets 68,000 votes from the rest of the state. His total would be 120,000…just 33 percent of the vote.
Scenario #2: This one is even more optimistic. Everything is the same as Scenario #1, but Edwards does even better among New Orleans voters – carrying the city by the same margin as Hillary Clinton. He leaves the city with approximately 65,000 votes. If he is still at 24 percent of the vote outside New Orleans, he ends up with a vote total of 133,000, just 37 percent.
Neither of those scenarios should be read as predictions. They are hypotheticals based on a set of assumptions, many of which could turn out differently tomorrow. The point is that even with New Orleans turnout at the high end of recent mayoral elections (close to 90,000 with 80,000 also voting for Treasurer) and statewide turnout at the lower bound of the Secretary of State’s prediction, Edwards still falls well short. In fact, even under the most optimistic scenario where he gets Clinton-level support in New Orleans, he would still need to get his vote share outside the city up to 41 percent to have a shot at a majority in the election.
Why my estimate of New Orleans turnout in the primary did not work.
Let this be a lesson in extrapolating from small and unusual samples. To estimate the impact of local elections on turnout in a statewide special election, I looked at differences in turnout rate across all parishes during the two most recent statewide special elections (2006 and 2010) between places with a local election on the ballot and places without a local election while controlling for time-constant characteristics of parishes that might affect turnout. On average, local elections boost turnout by three percentage points in (two) statewide special elections. The problem with this approach is that there were no cases in which New Orleans held a mayor’s election at the same time as a statewide special election. In my defense, I wanted to make sure to focus on statewide special elections to make sure I was not extrapolating from regular elections to special elections. Also, there were instances of local elections in New Orleans during these elections, but not a mayor’s race. If mayor’s elections are a bigger than average draw for voters in New Orleans than in the rest of the state, then the average effect captured in my estimate is insufficient to make inferences about New Orleans.