One reason for this is that as Oxford university's Simon Wren-Lewis says: "The important costs of Brexit were always long term"; less free trade would mean weaker growth not just in exports but also productivity.
But there's another reason. It's that there's a big distinction between a good theory and precise quantification of that theory.
We've good reason to suppose that uncertainty is bad for capital spending. This is because having potential investment projects is like having a call option on equities: you often have a choice of exercising them now or later. The greater is uncertainty, the greater is your incentive to hold on to those options because they might become even more valuable later. This theory has meant that, for years, politicians of all parties strived - albeit with mixed results - to reduce uncertainty. Lady Thatcher said: "An economy will work best when it is built on a framework of clear and predictable rules on which individuals and companies can depend when making their own plans." And Gordon Brown spoke endlessly of the need for stability, of "no return to boom and bust".
This theory applies even if you are optimistic about the long-term effects of Brexit. Let's say this will lead to freer trade. Why should a company invest in the hope of this now, when it can wait and invest in the certainty of it later?
Theory, then, is clear - which is unusual. What's not so clear is the precise quantification of it. We've never left the EU before, so we've no precedent for gauging the effect of this sort of uncertainty. This problem is magnified by the fact that, traditionally, downturns are about what happens to a minority of companies. The vast majority of companies don't export, so the question is: by how much will a few companies respond to uncertainty? Such a question can't be answered precisely.
It is, however, reasonable to suppose that sterling's fall might not fully offset the dampening effect of uncertainty. This is because its impact in boosting exports has been weak in the past. We know this not just from econometric evidence (the Office for Budget Responsibility estimates that a 10 per cent fall in relative prices raises exports by only 4.1 per cent after two years) but from casual observation: sterling's slump in 2008-09 did only little to support exports.
And this effect might well be mitigated by the fact that a weaker pound - insofar as it is passed on to customers which won't wholly be the case - will raise import prices and cut spending power. The Confederation of British Industry's Anna Leach recently warned that this will lead to lower consumer spending. In fact, the recent strength of retail sales might be consistent with this: customers are buying goods to beat possible price rises.
But there's something else. Since June, we've learned that Brexit won't happen for a few years. This has reduced uncertainty about shorter-duration assets, which means companies can safely buy these: they can invest in, say, software or machine parts that'll be obsolete or worn out in two or three years anyway. But it means uncertainty does affect longer-lasting assets such as buildings.
And this is exactly what's happening. Purchasing managers report that orders have risen in manufacturing, but are still falling in construction.
The fact that economists are revising up their near-term projections does not mean the basic economics of Brexit was wrong. All it tells us is that attempts at precise unconditional forecasts often bring the economics profession into disrepute. But then, we've known this for years.
