The Median Voter Model says that democratic competition impels politicians to offer whatever policies maximally satisfy the median voter. Actual democracies have never met its highly restrictive assumptions: There is more than one election, fought over more than one issue, and in any case, nothing forces politicians to fulfill their campaign promises. Yet for decades, this Median Voter Model has been extremely useful — much more useful than these restrictive assumptions would have you believe.
In particular, if you frame policies as “Government should do more X,” “Government should do less X,” and “Government should stay the course,” public opinion data usually reveals that the latter is the median position. And the main counterexample, category-specific spending, usually goes away if you add a reminder along the lines of “Remember, more spending requires more taxes.”
But U.S. politics has grown so weird during the last decade that I suspect a revision to the Median Voter Model is in order. I dub this the Neo-Median Voter Model. It works roughly like this:
In primary elections, both parties myopically select their candidates. Partisan voters in both parties pay almost no attention to the impending general election; they vote expressively for whichever candidate most appeals to them.
What determines appeal? While policy positions play a role, charisma and name recognition matter more. But again, due to myopia this is not charisma in the eyes of the median voter in the general election. It is charisma in the eyes of the median primary voter.
This charisma is extremely scarce. A party is pretty lucky even to have one presidential candidate who seems charismatic to party loyalists. Candidates who actually seem charismatic to the median voter are rarer yet. Since 1980, Reagan, Clinton, and Obama had general charisma. Trump has partisan charisma only. The last election where both parties had a charismatic candidate was 1964, where Johnson (general charisma) clobbered Goldwater (partisan charisma only).
After the primaries, both winning candidates make a moderate effort to pretend they’re more mainstream than they really are. But both sides primarily play to their base in the general election.
Given this odd candidate selection process, the Median Voter Model still determines the election. The main underlying factors: per Achen and Bartels, economic conditions and peace or lack thereof immediately prior to the election; charisma; and which party’s primary voters’ issue positions are further from the median voter. In that order.
Caveat: Voters have a longer memory for inflation than for other economic variables, and they hate inflation.
Once a party gains the presidency, they pay much more attention to their base than to the median voter. And the presidency is the main thing that matters for policy. The non-delegation doctrine notwithstanding, Congress has long since delegated most of its power to executive agencies — and presidents have usurped most of the remaining power. The budget is the only notable exception.
Thus, it is only a slight exaggeration to say that despite much rhetoric to the contrary, the president is more than halfway to being a four-year dictator. Congress has trouble merely surmounting the filibuster; an actual veto override is almost impossible. Judicial oversight isn’t completely irrelevant, but the system allows presidents to keep imposing the same illegal policies using different legal rationales. Since each Supreme Court challenge takes about a year, you only need three bogus rationalizations to maintain a blatantly illegal policy for your whole term. See Biden on student loans or Trump on tariffs.
After three years in power, the next presidential election cycle starts. The primaries myopically select two candidates, and the mechanisms of the Neo-Median Voter Model once again kick in.
Punchline: The traditional Median Voter Model’s central erroneous prediction is platform convergence. In the Neo-Median Voter Model, platforms diverge widely, especially on issues where the base is passionate and the median voter in the general election is apathetic. And it’s amazing how many issues most people are apathetic about. The median voter is pro-choice even in most Republican states, but very few Republicans are going to vote for Democrats over abortion. They can be annoyed for a week or a month over pro-life policies, but not a year, and definitely not permanently.
Is the Neo-Median Voter Model bad news? Probably. It means that the famed checks and balances of the U.S. Constitution are now mostly impotent. But the really bad news comes from studying primary voters. The Democratic Socialists of America and the national conservatives are tiny shares of the population. But they intellectualize the economically illiterate cruelty of their respective party faithful. Never in my lifetime has there been a better time to hate politics.
But love it or hate it, the Neo-Median Voter Model is a good guide to modern American politics. If you know how to make it better, please tell me in the comments.
You say "Voters have a longer memory for inflation than for other economic variables, and they hate inflation."
It's probably more of a price level effect than an inflation effect
I'm confused - wouldn't basic game theory say that the instant you get to the second-stage general election that you immediately pander to the new median voter who is more moderate than the median voter in the primary? It seems like we need some rationale here.
How about: "primaries pull candidates away from the general-election median to pander to the partisan median voter. The political commitments (vague I know) required to win the primary then make it costly (although still possible) for nominees to move back toward the center afterward." That explains why we might observe token rather than full moderation in the general election.