It is taken as a general rule in political commentary that after one party holds the presidency for eight years, it is very hard for its nominee to make it three in a row. Only one party, after all, has done it in the last 60 years. The Democrats lost in 1968, Republicans lost in 1976, Democrats lost in 2000; only the Republicans won in 1988. There are examples of this rule holding in states that are politically marginal: No party held the governorship of Ohio for more than eight years from the 1840s until 1998, when Republican Bob Taft was elected governor to succeed eight-year incumbent Republican George Voinovich. And perhaps this is the exception that proves the rule. Taft, re-elected in 2002, became very unpopular in his second term, and Democrat Ted Strickland was elected by a wide margin in 2006, with Democrats sweeping all the statewide downballot offices that Republicans had held since the 1990s.
But is the rule so ironclad at the presidential level? At the terrific Gay Patriot blog, Daniel Blatt reminds us of something that I had not focused on, that in the eighth year in which one party held the presidency in recent years, the out-party candidate had a huge advantage in May and June polls—and either didn't win in November or came close to losing. And I note that in vivid contrast to that, putative Democratic nominee Barack Obama leads presumptive Republican nominee John McCain by exactly 1 percent.
Blatt's examples are 1976, 1988, and 2000. All are different, in important respects, from 2008. In 1976, party identification worked even more powerfully against the incumbent Republicans than it does today, and if not the Watergate scandal then the Nixon pardon worked more powerfully against the Republican nominee, incumbent Gerald Ford, than the Iraq war and congressional scandals and earmarks work against John McCain, who urged the winning surge strategy on Iraq several years before George W. Bush adopted it and who had nothing to do with the congressional scandals and earmarks.
As for 1988, I was puzzled then and remain puzzled now why Michael Dukakis had poll leads over George H. W. Bush as large as 17 percent (after the Democratic National Convention). Ronald Reagan's job rating was around 50 percent, 20 percent higher than George W. Bush's today, though 20 percent lower than Reagan's before the Iran-contra scandal broke right after the 1986 off-year elections. The economy was in good shape, despite the stock market crash of October 1987. The voting public knew little about Dukakis—why were they willing to put him in the White House by such large margins?
In 2000, the Democrats had a problem. Incumbent President Bill Clinton's professional job rating was very high, in the realm of 70 percent when he was threatened with impeachment, but his personal favorable/unfavorable ratings were very low, in the vicinity of 30 percent positive in the same period. I have taken this since as a retrospective vote in favor of the 22nd Amendment, which limits the president to two terms in office: Voters were saying, in effect, that having elected Clinton twice, they wanted him to stay there, however unsavory his personal conduct, especially since they did not have the option of granting him a third term. This they had granted only once to a president, to Franklin Roosevelt at a moment of existential crisis for western civilization and at a time when his Republican opponents had exactly zero foreign policy experience. Would I have voted for a third term for Roosevelt in 1940? Absolutely, just as Reagan, who was an admirer of Roosevelt all his life, did. And with the knowledge that his inexperienced Republican opponent, Wendell Willkie, was just about as good as Roosevelt on standing up to totalitarianism (see Charles Peters's wonderful book Five Days in Philadelphia, which I reviewed for the Wall Street Journal.)
The rule that a party has a hard time winning a third presidential term is one of those political science rules that seem less ironclad after close inspection. Hubert Humphrey nearly won the popular vote in 1968 despite the debacle of the Democratic Party that year (although, noting the Democrats' decline from 61 percent in 1964 to 43 percent in 1968, Humphrey's intraparty opponent Eugene McCarthy said he would take credit for the last drop of 1 percent if Humphrey and Johnson had taken credit for the other 17 percent). Gerald Ford would have been elected (without winning the popular vote) if he had gotten about 12,000 more votes in Ohio in 1976. Al Gore did win the popular vote nationally in 2000, even if you believe as I do that Florida did indeed vote for George W. Bush; 1,000 or so votes the other way there, and Gore would have been president.
That leaves the only clear case of a third-term victory for a president's party as 1988. Which in retrospect seems overdetermined: Reagan's job rating was relatively high, the economy was not in perceptible trouble, Dukakis was a cold fish. But the other examples show that the rule that a third term for the incumbent party is impossible is actually a close-run thing. John McCain faces headwinds in George W. Bush's low job rating, but not some historically inexorable rule. And, as Daniel Blatt's insight suggests, those headwinds may not be so strong as conventional wisdom suggests. Obama against McCain isn't where Dukakis was against Bush 20 years ago. Interesting, eh?