since in both cases a healthy electoral-vote margin substantiated the popular vote.

Probably the most famous close election was Wilson over Hughes in 1916. Even then, however, Wilson ended up with a popular-vote margin of almost 7 per cent, though for once the electoral-college vote was so close that a single state, California, which Hughes was expected to carry and Wilson won by a whisker, made the ultimate difference.

In popular-vote terms the two closest elections have both been recent: Kennedy over Nixon in 1960 (a margin of less than one half of i per cent) and Nixon over Humphrey in 1968 (less than 2 per cent). Owing to third-party candidates both victors entered office as minority Presidents, and Nixon’s 43.4 per cent of the popular vote in 1968 was one of the lowest ever recorded for a winner.

After this unimpressive showing to have won re-election in one of the greatest of United States Presidential landslides appears a tremendous achievement, especially for a candidate notably lacking in personal popularity. In historical perspective, however, Nixon’s triumph of last November seems somewhat less remarkable.

In seven of the last twenty elections the voters have been faced with essentially the same choice as in 1972: an incumbent President running against the candidate of the party that four years earlier had lost the White House. In all seven instances the incumbent President won: 1900 (McKinley over Bryan for the second time), 1916 (Wilson over Hughes), 1924 (Coolidge over Davis), 1936 (Franklin Roosevelt over Landon), 1956 (Eisenhower over Stevenson), 1964 (Johnson over Gold water), as well as 1972.

Not only did the incumbents all win, but six of the seven increased their party’s popular-vote margin over what it had been in winning the White House four years previously—excluding only Wilson, in the squeaker election of 1916. And five of the seven victories, the exceptions being 1900 and 1916, were of landslide proportions.

Added evidence of the electoral strength of incumbents comes from their overall record. Sitting Presidents have run to succeed themselves in thirteen of the last twenty elections (the seven already referred to, plus 1904, 1912, 1932, 1940, 1944, and 1948). They have won eleven times and lost only twice.

Furthermore, each of the two losses occurred in highly extraordinary, even unique, circumstances. In 1912 the Republican Party was blown in two by a revolt against the incumbent President, Taft, led by his predecessor and one-time patron, Theodore Roosevelt. With Roosevelt running as a third-party candidate, the rivals divided the majority Republican vote, which permitted the Democrat, Wilson, to win with fewer popular votes (41.8 per cent) than any candidate had received since Lincoln’s 39.8 per cent in the four-candidate donnybrook of 1860, on the eve of the Civil War. (Nevertheless, Wilson’s electoral-vote margin was conclusive: more than 4 to 1.)

The only other loss by a Presidential incumbent was Hoover’s to Franklin Roosevelt in 1932, when the country was in the trough of

the Great Depression and “Hoover” had become to millions almost a dirty word. The deeper significance of Hoover’s defeat, however, came from what happened afterward as Roosevelt, combining an unexpected capacity for innovative government with vast political skill, succeeded in securing a permanent hold on millions of the voters he had drawn away from their previous Republican allegiance and made the Democrats into the country’s majority party—probably the most significant single political occurrence in the United States since Appomattox.

The factor of party affiliation appears to be almost as important as incumbency in determining the outcome of a Presidential election. In the nine elections from 1896 through 1928, during the era of Republican ascendancy, Republican candidates won seven times, losing only in 1912, the year of party fratricide, and 1916, when the incumbent Wilson was narrowly re-elected.

In 1932, as we have seen, the Democrats became the majority party, and in the ten elections since (1936 through 1972) their candidates have won six times. Once again, however, extraordinary circumstances prevailed when they didn’t. In 1952 the immense stature and popularity of Eisenhower, combined with the onus of an unpopular war in Korea, turned the tables; and in 1956 an incumbent Eisenhower was easily re-elected. Twelve years later, in 1968, an unpopular war in Vietnam and the most successful third-party campaign since Theodore Roosevelt’s, by George Wallace, again hung an albatross around the Democrats’ neck. Even so, Humphrey almost beat Nixon—though he came much closer in popular vote than in electoral vote.

Thus, from the evidence of twenty sets of election returns covering a period of eighty years, it seems that American Presidents are elected according to a few very simple rules:

1. 1. If an incumbent is running, he will win regardless of party.
2. 2. If an incumbent is not running, the candidate of the majority party will win. These rules have failed to apply only four times in the last twenty elections: in 1912, 1932, 1952, and 1968. From these few exceptions a final rule emerges:
3. 3. The first two rules will fail to apply only when an election is influenced by a disaster of enormous proportions: an unpopular war, a severe depression, or a party torn in two. In such cases those whom the voters consider responsible will be penalized and their opponents rewarded regardless of incumbency or party.

Seen in perspective, then—and despite the mighty events and massive changes that have affected the United States over the last eight decades—the pattern of Presidential election results has remained remarkably consistent. Is this happenstance, or buried somewhere in the election statistics may there be a reason why?

Consider the following:

The two most populous states, California and New York, elect their governors in non-Presidential years; in both states the governorship race draws more votes than any other statewide contest except the Presidential. In 1966 in California 6,503,000 votes were cast for governor; in 1970, 6,510,000. But in 1968 Californians

cast 7,252,000 votes for the various candidates for President—11 per cent more than either governorship race attracted. In New York it was the same story: 6,006,000 votes for governor in 1966 and 6,095,000 in 1970, but 6,745,000 votes for President in 1968—again, about 11 per cent more.•

• Figures for the 1972 Presidential election are omitted as not comparable because they include the new eighteen-year-old vote and the recent dramatic rise in California’s population.

From these and similar findings in other states it is clear that Presidential elections regularly attract the votes of millions of people who simply don’t vote in other elections. These purely Presidential voters probably amount to around 10 per cent of all those who vote for a President—perhaps as many as 7,500,000 in 1968, for instance.

Ten per cent of the electorate is a lot. If a large proportion of this ioper cent were to vote one way, they could easily determine the result in a close election or turn a close race into a walkaway. Normally, of course, it could be assumed that no such thing would happen and that any group of voters numbering in the millions will divide up pretty much like everybody else. But purely Presidential voters, it turns out, aren’t like everybody else.

Extensive studies of voting outlook and behavior conducted at the University of Michigan coincide closely with the opinions of experienced politicians about this group: in comparison with regular voters, purely Presidential voters are less interested in politics, government, and allied matters, are less well informed about political issues and candidates, and are much less inclined to hold strong opinions about them. This does not mean that the Presidential voter is a political independent. Like most regular voters, he usually thinks of himself as basically a Republican or a Democrat—but he is seldom a strong partisan.

There is no way of knowing for sure how Presidential voters actually vote, but it seems quite likely that people of this kind—not much interested in political matters, not very clear on what the issues are or where the candidates stand, not especially concerned about the whole thing anyway but drawn to participate with everybody else in America’s biggest event — would be strongly inclined to vote for an incumbent President, a man whom they know who has demonstrated that he can run the country, rather than try some new fellow. Or if both candidates are new fellows, to vote for the one who belongs to “my party.” And, finally, to depart from this rather direct and simple line of thinking only when it’s overwhelmingly obvious that the ship of state has gone on the rocks and drastic action is called for.

If this reasoning is sound, what we are left with is a paradox indeed. Despite the millions of dollars and billions of words poured out across the land each four years to influence the election of a President, the decisive role in choosing the

winner may well be played by an amorphous, anonymous collection of citizens whose chief claim to distinctiveness is that by conventional standards they appear to be the least qualified voters of all and who probably care less than anybody else about how it all comes out.