The NIESR claims that Johnson's Brexit deal "would ultimately lead the UK economy to be about 3.5% smaller than it would have been had the UK remained in the EU." Reaction to this, I fear, highlights an important way in which the media and public misunderstand economics. They pay too much attention to numbers (and seek a spurious precision in them) and too little to mechanisms.

To see what I mean, consider the NIESR's reasoning. This starts from the fact that leaving the EU's single market will throw a little sand into the wheels of trade by imposing trade barriers - "not the classic barriers of tariffs, but the insidious ones of differing national standards, various restrictions on the provision of services, exclusion of foreign firms from public contracts." (I'm quoting Lady Thatcher, who called the single market "a fantastic prospect for our industry and commerce.") A few days ago, Dominic Raab and Priti Patel claimed that Johnson's deal was great for Northern Ireland because it gave the region "frictionless trade" with the EU. But if frictionless trade is good for them, its absence must be bad for us.

Now, you might object that these barriers will be small - and they will initially given that we start from the same regulations and product standards. But this is where another mechanism becomes absolutely crucial. My chart hints at what this is. It plots annual growth in world trade (as measured by the CPB) against UK GDP per worker-hour: the data starts in 1997 as this is when monthly GDP data begins. It's obvious that there's a strong correlation between the two.

Of course, a big reason for this is that the same things drive both, such as fluctuations in credit conditions, animal spirits and aggregate demand. But this is not the whole story. We've strong reasons to think that world trade drives productivity. One reason for this was pointed out in the first lines of the first book on modern economics. "The greatest improvement in the productive powers of labour" wrote Adam Smith "seem to have been the effects of the division of labour." Greater world trade means greater division of labour and hence more efficiency. Also, freer trade encourages exporting firms to expand, which raises productivity as workers shift to these better firms. And then there's the fact that more trade increases competition which spurs firms to raise efficiency and innovate. There's also a knowledge diffusion channel; if managers visit suppliers and customers overseas, they should learn from them ways of increasing efficiency. As Danny Blanchflower says, you can learn a lot just by walking around.

These mechanisms are crucial. Over the long-run, they act as a multiplier. They cause apparently small barriers to trade to lead to significant losses of income because less trade means less productivity.

But we don't know precisely how powerful these mechanisms are. Yes, fact and theory tell us their direction: the natural experiment of dividing Korea proves that openness creates prosperity. But their magnitude is uncertain; this is why the NIESR's estimate differs from that of the UK in a Changing Europe (pdf).

In fact, there's another uncertainty. We don't know what future trading arrangements will be. Brexiters claim the frictions in trading with the EU will be minimized by a deep free trade agreement, and compensated for by freer trade elsewhere. Many, though, are sceptical of this: as the old saying goes, a bird in the hand is worth two in the bush. Even 10% rises in goods exports to China or Japan would not compensate for a mere 1% loss of exports to the EU - and 10% is perhaps an optimistic estimate of what free trade deals typically achieve.

For these reasons, Julian Jessop is right to say there are big uncertainties around the NIESR's estimate. This does NOT, however, justify philistine sniping at economists: "what do they know, ner, ner?" Instead, it directs our attention to the main issues: how strong is the link between trade and productivity? How likely are we to get meaningful trade deals? And what would these do to actually boost trade? It is possible to have a rational, intelligent debate about these issues - not that we'll get one on the BBC.

Julian says these uncertainties cause him to "give little weight to long-term forecasts." Some of us, however, draw the opposite inference: what is the point of introducing such uncertainties into the economy when there was no pressing need to do so?