The paper asks:
Why, whether figures are good or bad, are economists and City analysts always taken by surprise?
Prediction is very difficult, especially if it’s about the future.
More seriously, the bits of economics that we’ve pretty much got right are in microeconomics. Rent control is a bad idea, demand curves slope downwards, that sort of thing. Macroeconomics well, there might not actually be any one single macroeconomic model that you could get everyone to agree upon. You can and do find Nobel Laureates vehemently insisting that the signature idea of another one is simply wrong.
And finally, think through what people are actually trying to do when predicting these sorts of numbers. Take, for example, unemployment. Roughly speaking 10% of all jobs are destroyed in the economy each year (so, 3 million of the 30 million total). Again, roughly speaking, 10% of all jobs are created anew each year. Again, roughly 3 million. The unemployment level, the one we’re trying to predict each month or quarter, is the balance of those two processes.
And what usually happens in a recession is not that more jobs are destroyed but that fewer are created, making that balance rise. So we’ve two flows, of some 3 million each way a year, and we’re trying to tell what the balance is going to be at any one point in time to 0.01% of the total number.
That’s just not easy. Thus the people who try to predict are often surprised.
Finally, it’s worth noting how weak many predictions are. Take the US job creation numbers. Last few months these have been 200,000, to 250,000 a month. But the error bar is enormous. They’re really saying that job creation has been between 100,000 and 350,000 (for the error bar is 100,000 either way). And no one is at all surprised by job creation being in that range.
Apologies, just couldn’t resist trying to get something sensible into at least one part of that paper, even if it’s only the comments.