RALEIGH – In politics, the quality of the candidates, the quantity of the resources devoted to their campaigns, and the content of the messages they convey make a difference.

Keen grasp of the obvious, huh? Well, don’t snicker at me. It turns out that in some corners of the academy, my opening statement would, indeed, be considered controversial. Somehow they convinced themselves that they could predict the outcome of political outcomes reliably without knowing anything other than key economic trends in the months or years leading up to the election. Particularly when it came to predicting presidential elections, these political scientists and economist claimed that what candidates said or did had only marginal effects, if any, on outcomes.

I certainly believe that there are many factors outside the control of candidates and campaigns that affect their fates. I also believe that political pros often overestimate the electoral impact of what they do, failing to account for the law of diminishing returns as well as how much effort it takes to move voter sentiment even a single percentage point.

But I never bought the “economic data explains all” theory of politics. And now there is even more reason to be skeptical, in the form of a new analysis by New York Times commentator Nate Silver, the creator of its popular FiveThirtyEight site.

Silver recently took a look at the most popular economic models for predicting presidential-election outcomes. Here’s what he found after examining the results for nearly 60 such models since the 1992 Bill Clinton victory over George H.W. Bush and Ross Perot:

Can political scientists “predict winners and losers with amazing accuracy long before the campaigns start”?

The answer to this question, at least since 1992, has been emphatically not. Some of their forecasts have been better than others, but their track record as a whole is very poor.

And the models that claim to be able to predict elections based solely on the fundamentals — that is, without looking to horserace factors like polls or approval ratings — have done especially badly. Many of these models claim to explain as much as 90 percent of the variance in election outcomes without looking at a single poll. In practice, they have had almost literally no predictive power, whether looked at individually or averaged together.

Ouch.

To those politicians and consultants chortling at the foolish conceits of the would-be academic prognosticators, I would urge less self-congratulatory chuckling and more self-reflection. Did you know that door-to-door canvassing is a much more cost-effective way to turn out your base than sending mailers, ordering robo-calls, or distributing mass emails? Did you know that after a certain point, the dollars you raise and spend to buy broadcast advertising may be doing your cause or candidate little good, particularly if your candidate is the incumbent?

It’s best not to approach the task of analyzing politics with the assumption that the task will ever be simple, can ever be captured in a single explanation or formula, and will be the same today as it was 10 years ago, or 10 years from now.

Human nature doesn’t change. But technology changes rapidly. Audience tastes, media-consumption patterns, and the partisan allegiances of various subgroups of the population change, too, albeit a bit more slowly. No, you can’t predict political outcomes based simply on last year’s unemployment rate or this year’s budget for TV ads. Politics wouldn’t be nearly as much fun if you could.

Hood is president of the John Locke Foundation.