A few weeks back, in looking at a recent study by SmartPolitics that showed a gradual hardening of statewide preferences in recent presidential election history,
I noted that this could be viewed through the lens of recent discussion about resurrecting the 50-state strategy (one of the most identifiable features of Howard Dean's tenure as head of the Democratic National Committee).
A piece of evidence that seemed to call for a renaissance for the 50-state strategy came from a 2013 retrospective by Louis Jacobson at Governing, which made the connection that the implementation of that strategy between 2005-09 also coincided with the best electoral years of recent vintage for the Democrats, particularly in states where they had not (in recent history, at least) been terribly strong.
With a presidential election on the horizon that is likely to be highly competitive, this week we dig a little deeper into the presidential end of the "50-state strategy success" thesis. Did the broader focus on states laid out by Dean as DNC chairman pay off tangible dividends in the one presidential election (2008) where it was in effect?
Follow me below for a look at the numbers, which as you will read, defy an easy explanation for whether such a strategy helped Barack Obama in 2008, or whether it will help the Democratic nominee in 2016.
To examine what impact, if any, the more inclusive strategy of the Dean years had on the outcome of the 2008 presidential election, I took advantage of some excellent groundwork laid by my colleague, Stephen Wolf. He put together a Partisan Voting Index (PVI) chart dating back to 1828.
(Even if you could not care less about any discussion of whether the 50-state strategy was key to an Obama win, the chart is very much worth your time.)
Let's start with some simple criteria. Dating back to 1988 (an arbitrary year, one chosen because all the data fit on the screen at once), I looked at how the PVI for 2008 ranked compared to the others, in terms of how favorable it was for the Democratic candidate.
This was in response to a point Jacobson made that I found lacking (and, to be fair to him, he also acknowledged as an imperfect measure):
Here's how the Democrats fared in the reddest of red states between January 2005 and January 2009, the period when the 50-state project was in operation:
Presidential performance: In 15 of the 20 states, the Democratic nominee saw an increase in vote share between 2004 and 2008. In three other states, the vote share remained constant. It dropped in only two states.
Jacobson also noted that in 2012, when the 50-state strategy had been allowed to wither on the vine, only two of those 20 states saw an improvement in Obama's performance.
Of course, the inherent flaw in that type of analysis is that Obama did considerably better across the board than Kerry did in 2004. Likewise, Obama 2012 saw drops in most states compared to Obama 2008, which is why his national vote percentage receded a couple of points.
So, raw vote percentages or margins are quite clearly an imperfect measure for determining the success of the 50-state strategy in 2008. A better measure would be to look at PVI, because that allows us to look state-by-state at where the improvements were made.
In theory, if the 50-state strategy was one of the guiding factors in the electoral landslide enjoyed by President Obama, we would see some of the states where Democrats usually don't devote a ton of resources enjoying their best performances in memory, even when measured against the PVI.
Alas, for fans of the 50-state strategy, when looking at presidential PVI in the years of study, we simply see no clear pattern. Aside from the fact that Obama did worse than ever in pockets of the South (but not the deep South, where hard-wired racial voting patterns have made the PVI in places like South Carolina, Mississippi, and Georgia essentially immovable), there is no obvious connective tissue between the states where Obama did markedly better than his peers, and the ones in which he tanked.
Consider the list of states where Obama did the best, relative to PVI, of any Democrat since 1988:
Colorado (tied for 1st)
Delaware
Illinois
Indiana
Michigan
Nevada
New Mexico
The only one on the list that is a reliably red state was Indiana. Perhaps not coincidentally, it was the Hoosier State that offered arguably the greatest upset of 2008, when the Democrats carried it for the first time since 1964.
Sharp historians of the '08 cycle, however, would be more than likely to suggest that, rather than the 50-state strategy, it was Indiana's placement late in a primary cycle (a cycle that had relevance to the very last day) that played a dominant role in the competitiveness of that state, as well as another Obama upset win in the cycle (North Carolina). After all, Obama had to put an outsized commitment into the state in May, which undoubtedly paid him dividends in November in traditional Democratic strongholds that saw a ton of attention just six months earlier.
And, indeed, virtually all of Obama's "best states" (when measured by comparative PVI) have a similarly plausible explanation. Illinois and Delaware, of course, were the home states of the members of the Democratic ticket. Michigan was famously abandoned by McCain late in the campaign, leaving Democrats to romp in a state that usually sees a ton of campaign cash. The other three states are all regionally lumped together in an area of rapid demographic change.
It is hard to point to a state and say, with absolute conviction, that the decision to "leave no state behind" was pivotal in its outcome.
In fairness, though, let's look at another measurement that could prove helpful. Let's do what Jacobson had done, and compare 2004 to 2008.
To do this, let's look at the simple shift in PVI in all fifty states (plus Washington, D.C.) from 2004 to 2008. In that, we do see some very logical movement, but we also see the first signs of a possible "50-state strategy" impact.
Consider the seven "biggest movers" in PVI from 2004 to 2008, in favor of the Democrats:
1. Hawaii (13 points movement in favor of the Democrats)
2. Indiana (6 points)
t3. Nebraska (5 points)
t3. North Dakota (5 points)
t5. Delaware (4 points)
t5. Montana (4 points)
t5. Utah (4 points)
The clear outlier here is Hawaii, which clearly had a "home state" impact in favor of President Obama. Furthermore, the gargantuan PVI shift is also owed to the fact the Republicans over-performed in Hawaii in 2004 (remember when the Bush-Cheney team famously put the state into play in the final weeks?).
Delaware also is a bit out on an island here, but this could be a combination of a "Biden effect" coupled with a gradual movement to Democrats in a state that at one point was seen as something of a bellwether.
The other four states, however, have a lot of common threads. They are all habitually red states, and they are all states where the President did considerably better than what would be perceived as "normal" for a Democrat. It would seem here that the additional infrastructure provided by an active and inclusive nationwide strategy might have generated legitimate dividends.
As it happens, of course, Obama only picked off Indiana and a single electoral vote from Nebraska's 2nd district (Nebraska, along with Maine, allocate electoral votes a bit more granularly, offering single electoral votes for the popular vote winner in each of the state's congressional districts). So, in the final analysis, while the "50-state strategy" might've padded the national popular vote numbers by making some traditionally red states closer than normal (North Dakota, you will recall, was a single-digit outcome in 2008), it had a fairly minimal impact on the final electoral vote totals in 2008. It may well have yielded Obama an additional twelve electoral votes, but little more than that, it would appear.
When one thinks about it, this actually makes sense. It is fair to assume that the goal of the 50-state strategy had precious little to do with presidential election outcomes.
After all, as Howard Dean said in an interview with David Dayen last year:
The 50-state strategy was never about giving the same amount of money to Alabama as you give to Colorado. Never about that. But it was about giving everybody a base, and some competence level to work off, and then they were on their own.
If the goal (as stated elsewhere in the Dayen interview) was to put a handful of staffers into each state, that is unlikely to make much of a dent in what will inevitably be a massively funded presidential election. But in a battle for a majority in the House of Representatives, or a battle for the balance of power in a state legislature, the "bang for the buck" is considerably higher.
So, in the final analysis, the inclusive strategy of the Dean years might not have gotten Obama elected president, but it is fair to conclude that presidential victories weren't anywhere close to the goal of the strategy. The advantage downstream seems difficult to deny, given the ability of Democrats in the 2006 and 2008 cycles to compete in states and localities where Democrats traditionally took a beating.
As we said last week, of course, there is a chicken-or-egg debate to be had: Democrats also have not often enjoyed the type of tailwinds they experienced in those two election cycles, when they had a deeply unpopular president to build a campaign around. But did a more inclusive strategy help to add a few miles-per-hour to those tailwinds? That's the debate the Democrats need to have when looking ahead to the next several election cycles.
