49
the public has clear if latent policy preferences, a strong and recurring structure to its
preferences in a given policy realm, but where the analyst should not expect to see these
reflected in practical politics in a regular and extended fashion. Sometimes, this is
because events of the day place an obvious priority on some alternative policy realm.
Other times, it is because the priority for this policy realm is so high that both candidates
ought to be expected to be offering roughly the same program within it.
The policy realm of national security is an excellent example from the postwar
era. Ironically, in the period when the Cold War—the postwar struggle with international
communism—was at its ‘hottest’, the late 1940s and 1950s, the realm did not shape
presidential voting. And why would we expect it to, since neither party was willing to be
the ‘soft option’? However, when the two parties diverged in the 1960s and early 1970s,
public positions gained voting relevance, strongly. When the parties diverged further, to
their most dissensual in the 1980s, the voting influence of public preference too was at its
strongest. And of course, when the Cold War ended without some immediate
replacement in the realm of international conflict, the relevance of public preferences for
national security plummeted.
That is one way to say that public preferences need not register in politics to
confirm their existence, and that they should not be expected to register in the same way
in every election to affirm their consistency. But a better way to say the same thing, we
think, is just to summarize the issue evolution of American politics in the postwar era—
the place of policy substance in the public mind during the half-century when it is
possible to have a comprehensive issue structure with consistent measures. That story is
strong, nuanced, sensible, and, as ever, profoundly open-ended.
Convention