2016 Forecast: Fog

Precious little, as it happens. This is the most open Republican nomination battle in a generation, if not more. The large number of candidates, the unpredictable behavior of Donald Trump, and the lack of a clear frontrunner all conspire to make this race as hard to predict as defeating Garry Kasparov in a game of three-dimensional chess. Time and again, predictions about the state of the race have been quickly exposed as erroneous.

These forecasting errors could persist. In fact, there are four very good reasons why we might not have our first clue about the real shape of the race until the votes are tallied in Iowa and New Hampshire. And as for getting a sense of who will actually win this thing, that could still be months away.

First, wild, late swings in the primary polls have been so frequent in recent cycles that we would be remiss not to anticipate something similar in 2016. Newt Gingrich had a lead late in the pre-primary polls of 2012. Barack Obama charged from behind in 2008, as did John McCain that year. John Kerry emerged as an easy winner in 2004 after being nowhere in the polls for most of 2003.

Why does this seem to happen again and again? Partly because these are primaries, and the usual party labels therefore tell us nothing. Partisan affiliation is an important store of information, offering voters a sense of where candidates stand on a whole host of issues. But in primaries, voters settle on a candidate without that valuable heuristic. On top of that, primary combatants rarely develop meaningful distinctions on key issues, because they are all on the same side of the ideological spectrum. Voters are left to make choices based on an impressionistic sense of qualities—character, likability, and electability.

Poring over the relative merits of the candidates is a fun way for a political junkie to spend a Saturday evening. But most voters—even most primary voters—do not have that kind of time to dedicate to politics. So they are prone to make their final choices relatively close to the day of the primary or caucus. What makes it tricky for pollsters is that few people are actually willing to respond “I don’t know” when asked which candidate they prefer. Instead, they’ll offer a provisional opinion that is subject to change as they consider their options further. The primary polls can thus swing wildly, even late in the cycle.

Second, political polls in general have been functioning less and less well, and nobody is entirely sure why. We saw this in the 2014 midterm, when the GOP largely outperformed its poll numbers on Election Day. Obama enjoyed a similar overperformance in 2012. And the phenomenon has also happened overseas—with the Conserv-atives in Britain and Likud in Israel doing better than the pollsters predicted in 2015.

Theories abound, but there are no clear-cut answers. It is important to appreciate that while the news is flooded with a lot of polls at the moment, quantity does not necessarily yield quality.

One troubling portent is that many pollsters are still doing polls of registered voters, rather than likely voters. At such a late date in the cycle, it seems the only reason for this is budgetary: Given the high nonresponse rate to political polls, a refined sample of likely voters is more time-consuming and costly to collect. Turnout in presidential elections, after hitting a low in 1996, has been on an upswing, so there is less of a difference between actual voters and registered voters than there might have been 20 years ago. But primary turnout remains just a small fraction of general-election turnout, so the polls right now include a lot of people who will not actually vote in the primary. Do these nonvoters have different opinions from actual voters? If they do, the polls could be off by a wide margin.

Third, as anybody who followed the Barack Obama-Hillary Clinton battle of 2008 knows, the primary process is a quest for convention delegates, not for votes as such. It is noteworthy that states coming early in this cycle are required to distribute delegates proportionally, which means narrow vote victories will yield only tiny delegate leads. Later on, however, states are free to allocate according to the principle of winner-take-all, so a small margin of victory in a large state could be hugely consequential in amassing delegates. It is easy to come up with scenarios that “game out” the process over the months ahead, but the large number of assumptions one has to plug into such a computation renders the results hopelessly speculative.

Fourth, in every cycle there is talk of momentum, and for good reason. It matters who is up, who is down, who has exceeded expectations, who has fallen short. Without disputing the importance of momentum, though, a relentless focus on it can cause one to miss the forest for the trees. The bigger picture is that the primaries are the process by which a party, broadly defined, selects a nominee. As Anthony Downs outlines in An Economic Theory of Democracy (1957), such contests are usually characterized by the “median voter theorem,” whereby the candidate who is closest to the middle of the electorate is the one who wins.

There are a number of important caveats that could be added to this theorem. That is especially the case for a primary system in which voters do not all vote on the same day. With sequential contests, the calculus changes greatly. Still, the median voter theorem is a handy rule of thumb: A party usually selects a nominee who is broadly acceptable to most constituencies.

This is quite helpful for understanding how momentum works. If a candidate whom most of the party likes—or at least does not dislike—wins an early contest, it is a decent bet that he or she will develop some momentum. Think of John Kerry after his victory in Iowa in 2004, George W. Bush after his South Carolina triumph in 2000, and Ronald Reagan after New Hampshire in 1980. The parties were basically content with these men as their candidates and were happy to follow the signals sent by the early states. If, on the other hand, a party is internally divided, momentum can stall as factions balk at accepting the choice of the other factions. Think of Gerald Ford in 1976, Jimmy Carter in 1980, and Barack Obama in 2008. All three had momentum at one point or another in the contests, but they could not sustain it in the face of lukewarm support (or outright opposition) from certain factions.

Which dynamic will govern this cycle remains to be seen. It is easy to envision somebody challenging Donald Trump through the entire cycle, even if the real estate mogul should break out early. Too many Republicans would be too unhappy with a Trump win to allow for overawing momentum. The same might be true of Ted Cruz, who has become a darling of the grassroots by infuriating many Republicans in government. Accepting Cruz may be a tough pill for them to swallow, and collectively they have the power at least to bankroll a persistent challenger.

These four points are undervalued, and sometimes overlooked altogether. Total them up, and we have a whole mess of uncertainty—more so than most analysts fully appreciate. In a month, we should have a better grasp of what might happen, but it could still be months before we know what comes of all this. For now we must be patient and acknowledge that there is just too much we do not know.

Jay Cost is a staff writer at The Weekly Standard and the author of A Republic No More: Big Government and the Rise of American Political Corruption.