One of economists' favorite things to say is: "Economics only makes predictions at the margin. We can't tell you how much people will buy, only that they'll buy less when the price goes up." Sometimes this is true. But there are plenty of cases where economics makes predictions about levels, not just marginal responses. And in truth, these are our most interesting predictions - even if they're not our most accurate.
Thus, there are economic theories predicting that people simply won't vote, join revolutions, or give money to charity. Why not? It doesn't pay. Similarly, there are economic theories predicting that firms will set price equal to average cost, that individuals will discount future payments at the rate of interest, and that no one will cooperate in a finitely-repeated Prisoners' Dilemma game.
Are any of these theories true? Some are close to true, anyway. Americans keep about 98% of their income for themselves, however much they say they care about the poor. But in any case, the existence of these theories shows that economists do more than just predict marginal effects. We also try to predict levels.
And it's really when we predict levels that the world takes notice. People learn less about politics when the marginal benefits fall? Yawn. People know next to nothing about politics because it pays next to nothing? Neat!
Why do economists perpetuate the marginal myth? The main reason, in my view, is that economists start by making bold level claims. In the face of criticism, however, they retreat to making mellow marginal claims - and during their retreat, they deny that they ever aspired to more. Talk about sour grapes.
What's an honest economist to do? Make level claims when they're basically true, and admit that other level predictions don't pan out.