A lot of the local press was poised on May 1 to portray this year’s special election as a replay of the 2010 special election. Ed Markey was cast as the second coming of Martha Coakley, the longtime Democratic officeholder with progressive positions but a less-than-scintillating public persona. Gabriel Gomez, of course, would be playing the Scott Brown role: the telegenic “moderate” and “bi-partisan” Republican, a “fresh face,” etc., etc.
Gomez wants to follow the same path to victory Brown followed in 2010. He is going to get most of the 13%, the registered Republicans in Massachusetts. But they’re not enough. His only hope is in cleaning up among the unenrolled, who make up half the state’s voters, and peeling off some registered Democrats who aren’t necessarily liberal. Like Scott Brown in 2010, Gomez is an empty slate. Like Mitt Romney, he’s more slippery than an eel when it comes to his policy positions.
His dubious easement deduction, calling Markey “pond scum,” out-of-step positions on assault weapons, and antagonistic relationship with the truth may have taken some of the luster away from Gomez, but there’s still a chance it might work. The early polls (with the exception of Suffolk’s, which seems to have oversampled liberal Democrats) indicate a tight race. As we learned in 2010, a tight race in a special election comes down to enthusiasm and turnout. And to maximize turnout we have some work to do. Read on for more numbers than you probably want to see.
Shortly after the Elizabeth Warren-Scott Brown Senate election in November, I took a look at the voting numbers. Two trends jumped out at me: (1) Turnout was very high, particularly in urban Democratic strongholds; (2) Warren outperformed Martha Coakley by a few points in most towns around the state, about 330 of the 351. In about half of the towns where Warren underperformed Coakley, the drop was less than one percentage point; only in one small Berkshire town did it exceed five percentage points, and the sample is small enough to make that statistically insignificant.
There are two possible primary explanations for Senator Warren’s having outperformed Coakley in most of the state. First, that she herself was a better candidate than Coakley, swinging some votes and spurring turnout among Democratic voters who stayed home in 2010. Second, that she benefited from high turnout in a Presidential year and Barack Obama’s coattails.
In my view there’s little doubt Warren was, despite taking a while to become comfortable on the stump, a better candidate in 2012 than Martha Coakley was in 2009-10. And Scott Brown, with his nasty attacks, was a worse candidate than the Scott Brown of 2009-10.
Nonetheless, looking at the numbers not only for the two Senate elections, but also for our last two gubernatorial elections and the 2008 and 2012 presidential elections, I believe turnout in a Presidential year was a much bigger factor in Warren’s nearly eight-point win than the respective virtues of the candidates.
The biggest shift from January 2010 to November 2012 was in the composition of the electorate. PPP, in its penultimate January 2010 poll, noted that “it looks like the electorate in Massachusetts will be considerably more conservative than the one that showed up in 2008. Obama took the state by 26 points then, but those planning to vote next week only report having voted for him by 16.” And so it was.
Many have ascribed low Democratic turnout in January 2010 to demoralization over the healthcare dickering in Washington and Coakley’s lackluster campaign. I believe those factors are overstated, for the simple reason that turnout in the 2010 special election was consistent with turnout in the typical non-Presidential election in Massachusetts. In fact, more votes were cast in that special election than in the 2006 race for Governor.
As the table indicates, the big difference in turnout is Presidential year vs. non-Presidential year. The numbers show, with remarkable consistency, that about 800,000 to 900,000 more voters will show up in a Presidential year, more than a 30% increase. There is no question this discrepancy hurts Democrats more than Republicans. On the graph below, the little blue marks represent the towns and cities in the Commonwealth. The x-axis represents the increase in votes cast in the 2010 Senate race compared to the 2010 special election. The y-axis represents the percentage of those “new” votes going to Elizabeth Warren (i.e. (Warren votes in a town – Coakley votes in a town)/((Warren votes in a town – Coakley votes in a town) + (Brown ’12 votes in a town – Brown ’10 votes in a town))).
The overwhelming majority of towns are above the 50% line. That means that, in most towns, Democrats benefited more than Republicans from the increased turnout in 2012. The line of best fit on the graph shows that, the higher the increase in turnout, the higher the share of the “new” votes Democrats received.
This is not surprising. It reflects the known fact that potential Democratic voters are most likely to stay home for non-Presidential elections. As you can see, in cities where turnout was up by over 80%, Democrats received from 60% to nearly 90% of the “new” votes. If we don’t turn those people out this time, we will be missing out on that sea of blue in the top half of the graph.
To win this election, we need to make sure our voters show up. So where do we focus our efforts? First, in the cities.
Urban Turnout in Recent Massachusetts Elections
Although turnout was higher everywhere across the state in 2012 than in January 2010, the difference is most stark in (mostly heavily Democratic) urban areas.
I looked at 16 cities in this category: Boston, Brockton, Cambridge, Chelsea, Fall River, Fitchburg, Holyoke, Lawrence, Lowell, Lynn, New Bedford, Pittsfield, Salem, Somerville, Springfield, Worcester. These cities, all of which Elizabeth Warren won in 2012, have a combined population of over 1.9 million people. They represent over 29% of the state’s population. They are chronically underrepresented in election results, due to a higher number of noncitizens and lower registration and turnout rates. Even in the 2012 election, in which we rightfully were proud of urban turnout, only 22.6% of the votes in the state were cast in these cities. And that’s a good year. In the 2010 special, those 16 cities accounted for only 19.2% of the statewide vote. The 2010 special election turnout in the 16 cities was lower, but only slightly lower, than usual in a non-Presidential year. Therein lies our challenge.
(We can quibble over which cities should have been included, but I picked these 16 because they have large concentrations of potential Democratic voters who often don’t turn out except for Presidential elections. Fitchburg, for example, is not the Democratic stronghold that, say, Cambridge is. I included it because it’s really a tale of two Fitchburgs. The city has many conservative-leaning voters who show up regularly. But in 2012, when turnout in Fitchburg was up 53.4% over the 2010 special, we turned a 60-40 Brown city into a city Elizabeth Warren narrowly won. We did this by more than doubling the number of Democratic voters while Brown’s vote total went up only 26%.)
I looked at six recent elections: the 2012 Warren-Brown and the 2008 and 2012 Presidential races (in Presidential election years), and the 2006 and 2010 Governor’s races, and the 2010 Senate special election (in non-Presidential election years). As the table shows, on average the number of votes cast in these 16 Democratic cities goes up 55.1% in a Presidential year. Each of them outpaced the statewide average gain in votes of 38.5% and far outpaced the average increase of 34.2% in the state outside these 16 cities. Moreover, the trend is most pronounced in the poorest of the 16 cities, which are most heavily Democratic.
There is a remarkable consistency: in both Presidential years, the 16 cities represented over 22% of statewide vote. In both Gubernatorial elections (still in November but Obama not on the ballot), that number fell to under 21%. In the 2010 special, the number fell below 19%, and there’s little historical reason to assume it would be much higher in another off-cycle special election unless we work hard to make it so.
Comparing the 2010 Special with the 2012 Senate Election
The difference between the 2010 special election (lowest urban turnout of the six elections I looked at) to the 2012 Senate election was especially dramatic. The number of votes cast in these 16 cities was 65% higher in 2012, when Brown lost, than in 2010, when he won. In the rest of the state, the number of votes cast was up by less than 35%.
Such high turnout meant Scott Brown got more votes from these 16 cities in 2012 than in 2010. But the increase in urban turnout favored Democrats much more than it favored Brown. In 2012 Brown received about 201,000 votes in the 16 cities combined, an increase of almost 50,000 over 2010. But Warren received a total of 502,922 votes in the 16 cities, almost doubling Coakley’s total of 273,490. As a result the Democratic share of the vote in those 16 cities rose from 63.3% to 71.5%. Put even more starkly, the number of Senate votes cast in those 16 cities rose by 277,255 from 2010 to 2012. The Democratic candidate got 229,432 of those additional votes: almost 83 percent.
As shown in the table below, there is a clear correlation between the increase in votes cast in a city and improvement in Democratic performance. It’s simple: the more voters we turn out, the better we do, even if the Republicans turn out more as well. We have more sporadic voters than they do.
In November 2012 we turned out a lot more urban Democrats who stayed home in 2010 than they did Republicans. And our share of the vote in those cities went up significantly. The difference in terms of raw votes was even greater. In the 16 cities, Warren had a net gain (compared to Coakley) of 181,609 votes over what Brown gained (compared to Brown ’10). Let’s not forget that her final margin of victory was 229,228 votes. Her net improvement in just those 16 cities accounted for nearly 80% of her margin of victory. Those 16 cities also accounted for almost a third of her total votes. Outside those 16 cities, Brown actually beat Warren by over 70,000 votes.
For the reasons stated above turnout in the cities is, in my view, the most important factor for Democrats. The biggest numbers of votes are to be had there. But we benefit from increased turnout everywhere else too: the rest of Warren’s margin came from doing just a little better in the rest of the state. There were towns where Warren underperformed Coakley, but not too many (about 20, think Wellesley or Lincoln) and not by much (mostly between 0-2 percentage points lower). Virtually everywhere in Massachusetts Warren’s performance was a few points better than Coakley’s. Even in towns she lost pretty badly.
Due to the turnout explosion, in every single town in Massachusetts, the number of Scott Brown votes went up. But in over 300 of the 351 towns and cities, Democratic votes went up more, as shown on the graph with the blue marks. Just after the election, John Walsh told the Globe that turnout was way up in his hometown of Abington, cutting into Brown’s margin there though Brown still won handily. Looking at my nice spreadsheet, I see that 32% more votes were cast in Abington this time compared to 2010, and that Warren’s share was 39.9%, compared to 33.4% for Coakley. Brown’s vote total went up 791, his opponent’s 1,195, a net gain of 404 for the Dems. We were 50% more effective (1,195/791) at getting new Democratic voters out than they were at getting new Brown voters out in Abington. Across 300+ towns, that adds up.
What We Need To Do
Can we do it again? It’s going to be tough. When you really look at the numbers, a lot of the factors cited for low 2010 turnout (snow, Coakley’s lackluster campaign) seem less compelling. More relevant is the simple fact that turnout in any non-Presidential election, but particularly an off-season special election, is always going to be lower than in a Presidential year, especially among Democratic voters in strongly Democratic cities.
To win this special election we must do all we can to get out the Democratic vote, especially in the 16 cities but really everywhere. That’s a truism, but it’s really urgent this time. We can’t count on 2012-level turnout happening without our very best efforts. Our other hope is to soften up Gomez’s favorability among Dems and independents. His tax deduction looks more and more fishy as time goes on, and suburban parents don’t like assault weapons. But really, folks, turnout. Turnout. Turnout. We have the voters, we just have to get out there to make sure they show up.
John Walsh has been making the rounds, telling people that after so many tough elections in recent years he knows we’re tired. But, he adds immediately, we should be more tired of Republican obstruction in the Senate, and of our ideas getting short shrift while the Tea Party dominates the national discourse. Electing Gabriel Gomez to the Senate would be a disaster. It would, as John says, cancel out all the great work we did getting Elizabeth Warren elected. It would leave Massachusetts “silent,” with a 1-1 vote on many key issues instead of a strong 2-0 in favor of progressive ideals. And – as Scott Brown’s election in January 2010 did – it would send a message to the nation and the mainstream media that the Democrats are down in the dumps again, that the GOP is ascendant, that the Benghazi and IRS nonsense have “harmed the Democratic brand,” blah blah blah blah blah.
Sending Ed Markey to join Elizabeth Warren in the Senate, on the other hand, would create a Dream Team. Markey may not be the most electric orator, but he’s been a very effective Congressman (even Norman Ornstein agrees) and is a strong voice on every issue dear to us. He is a leader on the most important issue we have – climate change – just as Warren is a leader on issue 1-A – breaking the Wall Street chokehold on our nation.
That’s why I’ll be out on the doors and phones. Because I really want to win this one, and I know we have built-in challenges in a special election. I know you want to win also, and I hope you’ll get out there too. The time is now. Memorial Day is over, we’re almost in June, we have less than four weeks. This is not the one to sit out.