Symbolically Conservative, Operationally Liberal

Two questions, two answers

Ask a national sample of Americans whether they consider themselves liberal, moderate, or conservative, and for decades more people have chosen "conservative" than "liberal." Then ask the same sample, program by program, whether the federal government should spend more, less, or the same on education, health care, aid to the poor, environmental protection, and Social Security. On most items, majorities want spending held steady or increased.

Taken together, those two results describe a public that prefers the conservative label and the liberal program. Christopher Ellis and James Stimson gave the pattern a name in Ideology in America (Cambridge University Press, 2012): Americans are "symbolically conservative" but "operationally liberal."

Symbolic ideology is the label a person claims and the images that go with it: self-reliance, tradition, limited government, the flag, the small town. Operational ideology is what a person actually wants government to do when confronted with concrete choices. Ellis and Stimson's contribution was to show that for a very large share of the electorate the two point in different directions, and that this is not a marginal quirk but one of the central facts of American public opinion.

The reasons they offer are instructive. The word "conservative" carries connotations in American culture well beyond public policy: prudence, religious seriousness, respect for established ways. Many people who favor a generous welfare state nonetheless think of themselves as conservative in that broader, dispositional sense, and the survey question gives them no way to say so except by choosing the label. Meanwhile "liberal" has for decades been used by its opponents as an epithet, which discourages people from claiming it even when their policy views fit. The label and the program drift apart because they are answering different questions.

Why the line cannot hold this

Here is where the finding connects to the argument of this series. A single left-right axis assumes that symbol and substance occupy the same place. If you are conservative, you want less government; if you want less government, you are conservative. The whole convenience of the line depends on that identity. Ellis and Stimson show that for millions of citizens the identity fails. Such a voter is at one point on the line by self-description and at a different point by preference, and there is no way to plot her once.

Consider what this does inside a two-team system. Campaign strategists know the symbolic numbers and the operational numbers both. One team therefore campaigns on symbols, limited government and traditional values, while avoiding specifics about which programs it would cut, because the cuts are unpopular with its own voters. The other team campaigns on programs while avoiding the label, because the label is unpopular with people who like the programs. Neither team is lying, exactly. Each is speaking to the half of the voter it can reach. The voter, in turn, ends up with a representative who was chosen on symbol and governs on substance, or the reverse, and feels, correctly, that something was lost in translation.

In my years as a chaplain I spent a great deal of time listening to people describe themselves, and I learned that self-description is rarely false but often incomplete. A soldier who called himself a traditionalist might also be the first to insist that the Army take care of a comrade's family. He saw no contradiction, and he was right not to. The contradiction appears only when an outside observer insists that "traditionalist" must sit at a particular point on a line that also measures welfare policy.

Ellis and Stimson's symbolically conservative, operationally liberal citizen is one of the groups this series will keep returning to: people whose combination of views has no obvious home in either major party's bundle. The two-team system handles them by courting one half of their identity each election cycle and hoping the other half stays quiet. For a long time that worked. Whether it still does is a question I will take up when the series reaches the realignments of the 1960s, the 1990s, and the 2010s.

Where the evidence pushes back

Several cautions are in order. First, the pattern is asymmetric: there are relatively few symbolically liberal, operationally conservative Americans, so the finding describes one kind of cross-pressure rather than a general scrambling of the line. Second, spending questions have a known tilt; most people favor more spending on most things when no cost is attached, and some scholars argue this inflates the "operationally liberal" count. Third, as partisan sorting has advanced, the gap has narrowed among the most engaged partisans, who increasingly bring their labels and their preferences into alignment in the way Matthew Levendusky's The Partisan Sort (2009) describes. The symbol-substance gap is real and large, but it is largest among the less attentive, which is exactly where the parties have the least incentive to look.

Why this matters for the ballot

A voter whose label and whose preferences point in different directions is poorly served by a ballot that offers only one candidate per label. Fair access for independents and third parties makes it possible for a candidate to run on the combination that voter actually holds, rather than on the half of it one team finds convenient.

— Michael W. Cunningham, Ph.D.

Cover image: "Balance scale IMGP9755" via Wikimedia Commons. See the file page for author and license.