Ryan Avent says that he thinks the "plausible range of economic outcomes is, at the moment, quite wide indeed", and wonders whether "the size of this conceivable range is actually reflective of potential outcomes, rather than simple ignorance":
If it seems like things could go really well or really poorly, is it because outcomes are very dependent on our actions or because we have no idea what’s going on? And I suppose that if you want to understand the different approaches to policy advocated by liberals relative to libertarians, that question, and its answer, is key.
The easy answer is that this is not "simple ignorance". Studies have shown repeatedly and convincingly that the range of possible outcomes is nearly always greater than you think, not smaller. Economically speaking, actual results regularly come out quite far away from even pretty near-term forecasts, which is one reason why asking economists (or journalists, for that matter) to predict when we’re going to come out of recession is an exercise in futility.
On the other hand, that doesn’t necessarily mean that outcomes are very dependent on our actions. What action, for instance, did we take to send the price of oil plunging by $100 a barrel in the space of a few short months? That’s the kind of thing which can only happen in a highly complex and therefore wholly unpredictable system.
So I don’t think of liberals’ policy approach as being deterministic — if we do this, then the economy will do that. Instead, I think of it as working the other way around: if the economy does this, then we should do that. Right now, the economy is tanking, and so we should apply a large dose of stimulus. We can’t predict with any accuracy what the results will be, but there’s a very high chance that they will be better than if we did nothing, which is the default libertarian position.