Not every scientist is working on a cure for cancer, but it's surely one of the big open questions in medical science. Does economics have a cure-for-cancer question, one that every economist would sell his soul to answer?
A thought experiment: there's an economics seminar or lecture about to take place, in an hour, at a location an hour away from you. You hear a rumor about the content of this seminar. What credible rumor would make you jump in the car and high-tail it to the scene?
Again, being that it's probably foolhardy to speculate on what the next "big idea" might be, this might be an unanswerable question. Nevertheless, if it's not possible to think of an example, we might have a problem on our hands. Is this question different in economics than in other fields?
Broadly, there are three things that might qualify. First would be the uncovering of a new piece of evidence, a Dead Sea scrolls-type discovery that would be vital for some purpose or another. What might that be in economics?
Second, and related, would be proving of some unproven result. Whether an unproven provable theory can logically exist in economics is debatable; in theoretical economics, we're always subject to the underdetermination problem that skews the definition of "proof"; the theories are always logically consistent from within, and nothing is provable from without. In empirical work, we face not just the problem of where the evidence for our groundbreaking proof comes from, but if it is conceivable that the evidence could realistically exist to make such a proof. There a "proof" could rest on clever data collection or a new econometric technique that can interpret the real-world data in a usefully new way.
Third, we might have some methodological development, perhaps of the kind that saw "information" formally brought into the methodological fold in the 1970s. Whether this could rightly be called a methodological breakthrough is again not clear: is it not just a new application of the existing methodology?
Could neuroeconomics, partway closing the underdetermination door, do the trick? Will the modeling of complex systems by powerful computers help? All I know is that fewer than 1% of economics seminars are genuinely interesting in and of themselves (that is, apart from the paper and to the terminally curious); what could boost that ratio?