Economist Writing Every Day

The Democratic party won more elections than was broadly anticipated last week, though results fell roughly within a standard deviation of the highest regarded forecasts. There were, however, some patterns worth noting. Election deniers seemed to have systematically underperformed expections. Conservative democrats seem to have overperformed expectations, sometimes radically so.

Putting aside forecasts based on polls, make no mistake: the Democrats should have been slaughtered in these elections. High inflation, high gas prices, a frozen housing market, a declining stock market? I’m not Ray Fair, but his inflation+income+incumbency model has been accurately forecasting elections for 50 years for a reason. If Democrats overperformed as massively as it appears they have, then the most likely explanation is that Republicans did something wrong.

First, as is tradition, let us render tribute unto the median voter theorem by whispering Anthony Downs name before taking a sip of coffee. Now, again as is tradition, let us cast mild aspersions upon those who declared the theory obsolete at best, comically reductive and coarse at worst, a relic of a less sophisticated era of social science. Yes, politics exists in more than just the single liberal-conservative dimension and we are all characterized by multi-dimensional preferences. But the liberal-conservative spectrum is the dominant political shorthand for a reason. If politics really were a played out on a single dimension with single-peaked preferences, then the candidate closest to the preferences of the median voter would always win. That has to count for something.

But that’s all just (mildly silly) social scientific gloating. What might we have learned, if anything, from the last two elections? What I think I’ve learned is that the elections– not the discourse, elections, are moving to the center relative to six years ago, and the forces behind it are the same ones that gave us Amazon and the golden age of television. The fixed costs of serving the long tail of consumer demands.

Quick refresher. Amazon succeeded as a retailer not by selling you the 90% of things that everyone else buys, but by offering the 10% of things that you want that relatively few other people want. Once you enter into the long tail of consumption, its extremely difficult for a brick-and-mortar to offer everyone what they want because the opportunity cost of stocking and shelving are to high relative to the small number they will sell. Similarly, in the age of 5 channels, the opportunity cost of niche entertainment is too high. You can’t make Mad Men for 2 million people when you have the potential to air a show that is close enough to the median viewer to grab the attention of 40 million. When the marginal cost of a product listing approaches zero, though, when a 500 channels cut audiences into thousands of slices, however, the math changes. Now the opportunity lies not in serving the lowest common denominator, but giving each person exactly what they want. That’s easier said than done, of course, just ask current Netflix shareholders.

And this appears to be exactly that is happening in the political discourse. The conversation around politics is becoming more niche and, yes, in many cases more extreme. Yes, we are increasingly living in echo chambers served by content producers more than happy to produce bespoke information bundles that will confirm your pre-existing beliefs at the highest possible pitch and volume. But elections are not the discourse. Elections are still played out in a one person, one vote construct and politics is still reducible to a single coarse dimension. Politicians have incentive to incite the id of voters, especially their base, to rile them up to turn out in greater numbers, but that is not without cost. Aligning yourself with your party’s id distances you from the true median voter, who through the forces pushing and pulling political brands will always find herself wary of the most extreme elements of clubs she isn’t particularly interested in joining.

Parties continue to exist and thrive for the same reason that Amazon does. Super-niche product sellers and echo chambers can thrive independently, but when the moment of aggregation arrives, scale matters. Where decision-making bottle necks, either with a credit card or a ballot box, there’s an enormous advantage to having a brand that lowers information costs and mitigates risk. Amazon wins because it’s boring. Amazon mitigates the risk of buying from a million niche producers, some of whom you might only buy from once in your life, but at the end of the day you know everything is almost definitely going to show up. Democrats won more last week than they should have because they’re boring. Voters knew the candidates, some of whom they may never vote for again, would show up to govern and maintain longstanding institutions. The Republican party, in their independent efforts to serve specific niches within their coaltition, let their collective brand deteriorate and, in doing so, failed to mitigate the risk facing the median voter.

Republican’s got caught up serving the discourse, fractured across a thousand channels, but elections aren’t carried out on a thousand channels. There’s still just one ballot box in every election. The median voter theoreom may be based on an unchanging analog model, but so is our democracy.