Steve Kornacki: What History Really Says About an '08 Landslide
I was trying to figure out where to lead things off today, and then I saw this Politico piece by David Paul Kuhn, which is already getting heavy play across the web. The main idea: Barack Obama will not win a landslide because history says that landslides are evident in polling by the end of the summer. Writes Kuhn:
In five of the six post-war landslides (defined as a victory of 10 percentage points or more) the eventual winner was ahead by at least 10 percentage points in the polls at the close of August, according to a Politico analysis of historical Gallup polls. Over the past week, however, Gallup’s daily tracking poll pegs Obama ahead of John McCain by a margin of 2 to 5 percentage points.
The one exception to the August rule was 1980. Ronald Reagan was trailing slightly in the August polls before surging forward to win by roughly a 10-point margin.
By comparison, the biggest post-war landslides—1964, 1972 and 1984—were signaled by a large, double-digit advantage held by the eventual winner at the close of August.
This is a deeply flawed and inaccurate analysis of history. The Obama-McCain race may end up staying close through Election Day – a scenario that seems more likely than not – but “the August rule” is a rather arbitrary invention with little meaningful history to support.
Start with the fact that public opinion polling was nowhere near as precise and pervasive in the first few decades of the post-war era as it is today. Had there been regular and reliable data in 1948, for instance, Harry Truman’s upset of Thomas Dewey would have been visible weeks earlier and the final outcome wouldn’t have produced a shock that’s reverberated for decades. So in assessing “the August rule,” we can dismiss the post-war elections of ’48, ’52, ’56, and ’60 – at least.
Then, we need to consider the nature of each individual election. Kuhn correctly notes that the landslide winners of the ’64, ’72 and ’84 elections all held wide leads in polls in late August. But there is a common thread between these races: they all featured popular incumbent presidents who were blessed with fractured and politically-suicidal opposition. Elections that involve incumbents are inherently different from those that don’t, in that they serve almost exclusively as a referendum on the president. In this sense, “the August rule” may have some limited application: A large lead for an incumbent at the end of the summer signals a generally content electorate, and hence a likely blowout victory for the incumbent.
Additionally, ’64, ’72 and ’84 each represent a particular and extraordinary type of landslide – the near-total wipeout – that no one (not even the most devout Obamaphile) has been forecasting for this year. Richard Nixon in ’72 and Ronald Reagan in ’84 each won all but one state, while Lyndon Johnson ravaged Barry Goldwater in every corner of the country in ’64, save for the heart of the old Confederacy, where Goldwater’s States’ Rights platform resonated. Those who are most bullish on Obama, by contrast, have been predicting a less thorough landslide – something more consistent with 1980, when swing voters broke heavily toward Reagan in the last week, leading to a 10-point, 44-state rout...
...So what year is 2008 like? It’s obviously no 1964, 1972 or 1984. But it could easily be a 1980, with Obama playing the Reagan role and trying to reassure voters who badly want to toss out the incumbent party. It could also be like ’76, with McCain as Gerald Ford, the likable and “tested” candidate who fed doubts about his opponent’s experience and prompted a surprising number of voters to overlook their hostility toward the incumbent party. In both of those years, what happened on Election Day could not have been reliably predicted from polling in August. It is entirely possible that this year will be the same.
Read entire article at New York Observer
In five of the six post-war landslides (defined as a victory of 10 percentage points or more) the eventual winner was ahead by at least 10 percentage points in the polls at the close of August, according to a Politico analysis of historical Gallup polls. Over the past week, however, Gallup’s daily tracking poll pegs Obama ahead of John McCain by a margin of 2 to 5 percentage points.
The one exception to the August rule was 1980. Ronald Reagan was trailing slightly in the August polls before surging forward to win by roughly a 10-point margin.
By comparison, the biggest post-war landslides—1964, 1972 and 1984—were signaled by a large, double-digit advantage held by the eventual winner at the close of August.
This is a deeply flawed and inaccurate analysis of history. The Obama-McCain race may end up staying close through Election Day – a scenario that seems more likely than not – but “the August rule” is a rather arbitrary invention with little meaningful history to support.
Start with the fact that public opinion polling was nowhere near as precise and pervasive in the first few decades of the post-war era as it is today. Had there been regular and reliable data in 1948, for instance, Harry Truman’s upset of Thomas Dewey would have been visible weeks earlier and the final outcome wouldn’t have produced a shock that’s reverberated for decades. So in assessing “the August rule,” we can dismiss the post-war elections of ’48, ’52, ’56, and ’60 – at least.
Then, we need to consider the nature of each individual election. Kuhn correctly notes that the landslide winners of the ’64, ’72 and ’84 elections all held wide leads in polls in late August. But there is a common thread between these races: they all featured popular incumbent presidents who were blessed with fractured and politically-suicidal opposition. Elections that involve incumbents are inherently different from those that don’t, in that they serve almost exclusively as a referendum on the president. In this sense, “the August rule” may have some limited application: A large lead for an incumbent at the end of the summer signals a generally content electorate, and hence a likely blowout victory for the incumbent.
Additionally, ’64, ’72 and ’84 each represent a particular and extraordinary type of landslide – the near-total wipeout – that no one (not even the most devout Obamaphile) has been forecasting for this year. Richard Nixon in ’72 and Ronald Reagan in ’84 each won all but one state, while Lyndon Johnson ravaged Barry Goldwater in every corner of the country in ’64, save for the heart of the old Confederacy, where Goldwater’s States’ Rights platform resonated. Those who are most bullish on Obama, by contrast, have been predicting a less thorough landslide – something more consistent with 1980, when swing voters broke heavily toward Reagan in the last week, leading to a 10-point, 44-state rout...
...So what year is 2008 like? It’s obviously no 1964, 1972 or 1984. But it could easily be a 1980, with Obama playing the Reagan role and trying to reassure voters who badly want to toss out the incumbent party. It could also be like ’76, with McCain as Gerald Ford, the likable and “tested” candidate who fed doubts about his opponent’s experience and prompted a surprising number of voters to overlook their hostility toward the incumbent party. In both of those years, what happened on Election Day could not have been reliably predicted from polling in August. It is entirely possible that this year will be the same.