(1) Gerald Ford's job approval ratings were never under 37%, according to Gallup. (And that 37% was in January and March 1975, when the recession was hitting the country hard, when the Republicans had just done poorly in the 1974 elections, and when memories of the Nixon pardon were still fresh.) In 1976, Ford's job approval ratings were never under 45%--not even in June 1976, when he was trailing Carter badly. (The fact that Ford's job approval rating was actually a net positive in June--45-40--should have suggested that Carter's huge lead was unlikely to hold up.)
http://www.gallup.com/poll/23995/gerald-ford-retrospective.aspx
(2) With regard to 1960, the real question is not whether Nixon or JFK won the popular vote, but whether a system that has four keys in Nixon's favor and nine against him has something wrong with it when the result was a virtual tie. I know that Lichtman is fond of saying that the system is a binary one, that it does not intend to predict the *margin* of a candidate's victory, but surely results like this suggest that even if the system has never actually been wrong, it *could* have gone wrong under other circumstances and might go wrong in the future. (Of course if Nixon had won a clear popular plurality in 1960, Lichtman, writing in the 1980's, would have devised a different key system!)
I know that it has been argued that in 1960 JFK's Catholicism made the race much closer than it would otherwise have been. But even if this is true (as it may well be, though of course to some extent religion helped as well as hurt JFK, though it probably hurt him more than it helped) it means that we sometimes have to take account of factors not found in the keys.
(3) But, you might say, even if anyone can after the fact devise a system to explain all previous elections, Lichtman's system has worked for all the elections *after* he promulgated the system. True enough, but note Jonathan Bernstein's observations:
".... What he has is a combination of things that are generally causes (such as the economy) of incumbent party success, things that are effects of incumbent party success (such as the incumbent winning renomination uncontested and third party challenges), and things that are arbitrary and dubious (such as whether the candidates have "charisma").
"It's not surprising that you can "predict" the winner with that batch of stuff. After all, while Lichtman's system has worked since he debuted it for the 1984 cycle, a much simpler system that predicts the incumbent party wins barring an election-year recession also successfully calls the winners from 1984 through 2008, at least if you count 2000 for Gore (as Lichtman does). That doesn't make Lichtman wrong as much as it just means his system isn't telling us much that we don't already know otherwise."
http://plainblogaboutpolitics.blogspot.com/2011/08/not-buying-keys.html
(In 1952 to be sure there was no recession, but there was a seemingly endless war...)
(4) A lot of Lichtman's keys are *correlated* with electoral success rather than necessarily a *cause* of it. Take incumbency. Undoubtedly, a party is more likely to win a presidential election when its candidate is an incumbent. But this does not necessarily indicate that incumbency is an advantage. Rather, maybe it is simply a matter that *usually* the fact that an incumbent is running for re-election means that he has only been in office for four years, and absent a war or recession voters are willing to give him a second chance. By 1952, however, Truman had been in office for more than seven years, and the Democratic Party for twenty, and time-for-a-change sentiment was mounting. This would have hurt *any* Demcorat, but there is certainly no reason to think it would have hurt Truman *less* than Stevenson--even if Truman had won the nomination with little opposition (which is altogether possible--after all, Hoover had little Republican opposition for renomination in 1932, largely because the Republicans realized they were doomed by his record no matter who they nominated).
(5) In addition, in 1992 Lichtman and his co-author actually disagreed whether the keys pointed to a Clinton or a Bush victory! One reason: "In 1992, Lichtman called the "recession" key, Key 5, as lost for Bush even though subsequent analysis indicated the recession had ended in the summer. "The perception was there during the campaign, and that's what counts," Lichtman says."
http://www.lionhrtpub.com/orms/orms-2-04/frwhitehouse.html But if we are going to go on "perceptions" then part of the attraction of the model--that it lets us get past polls to more "objective" indicators--is lost.
In short, while Lichtman's system does point to some factors related to success in presidential elections, I wouldn't take it too religiously. If the system would point to a Truman victory over Taft in 1952, I would say that the system would be wrong.