"Globalisation, as we once knew it, is dead" said Rachel Reeves this week. Which is a problem for anybody wanting a high-wage high-productivity economy.

I say so for a simple reason: globalization in the sense of faster world trade growth raises productivity.

My chart shows the point, plotting three-year growth in world trade (as measured by the CPB) against UK productivity growth. It's clear that there's a correlation between the two. Of course, correlation is not causality. Some of the link is because the same things that depress one also depress the other - such as the financial crisis and the pandemic. But those events are not the whole story. The pre-crisis slowdown in world trade was associated with slower productivity growth; both grew weakly in the mid-10s; and both picked up in the late 10s.

In fact, we have strong reasons to think that faster world trade growth is indeed one cause of productivity growth (of many!). Productivity improvements "seem to have been the effects of the division of labour" wrote Adam Smith in the first sentence of the first book of modern economics. If you were to emulate Tom Good and try to become self-sufficient, you'd be poorer. What's true of individuals is true of countries: just look at North Korea for an extreme example.

What's more, stronger world trade means greater competition. That encourages firms to increase efficiency or to innovate more, both of which raise productivity. It also means that managers visit customers and suppliers overseas and so can more easily learn tricks of the trade from best practice overseas.

Globalization doesn't just raise real incomes by raising productivity, though. It also does so by cutting prices thereby making people's wages go further. The deglobalization caused by Brexit, for example, has greatly added (pdf) to UK food bills.

If globalization is dead, therefore, we face a big obstacle to raising productivity and real wages. Which makes it all the more important to look for other methods of doing so.

But, but, but. Reeves is right to say that globalization had a downside: "whole industries were allowed to decline or disappear."

This, however, is what productivity growth is. Productivity does not rise because we become better at doing the jobs we are doing. It does so because low-productivity jobs disappear and are replaced by higher-productivity ones.

A famous paper by Jonathan Haskel and colleagues showed this (pdf). They estimate that during Thatcher's great shake-up of manufacturing "between 1980 and 1992 single establishment firms experienced no productivity growth among survivors." Instead, all the substantial growth there was came because inefficient firms closed, more efficient ones opened, and because big companies shifted output from their less efficient sites to more efficient ones. A similar thing, Haskel has found (pdf), is true in retailing: most productivity growth is due to the closure of old stores and opening of new ones, rather than to incumbents upping their game.

Growth, as Schumpeter said, is a process of creative destruction.

Which means there is a tension between Reeves' two objectives. An economy of high productivity growth is one in which there will be insecurity as low-productivity jobs will be in danger of being destroyed.

Policy-makers can in principle ameliorate this trade-off by having a more generous welfare state to support those whose jobs are destroyed and more active labour market policies to help the unemployed into those better jobs. (Yes, I mean help, not bullying). But Reeves' speech gave no hint of thinking along these lines.

Instead, she emphasises the upside of better jobs:

That means cyber amongst the former coal and steel communities of South Wales, robotics and AI in business hubs in Lancashire's mill-towns and carbon capture and storage technologies pioneered and applied in the manufacturing landscape of Teesside.

But this still leaves the problem of job destruction. If these good jobs are created, people will move into them from low-wage work. Who then will do those jobs - the care work, the cleaning, the shelf-stacking?

This wouldn't be a problem if we had lots of unemployment or if Labour were endorsing high immigration. But neither is the case: Labour "want and expect net immigration to reduce."

Nor would it be a problem if the low-wage jobs were ones that we didn't want. But this is not so. "Skill" is an ideological construct: fund managers are considered skilled whilst care-workers are not. The "vibrant market" that Reeves wants might therefore move workers away from jobs that need doing.

Of course, the solution to this is to pay care workers more. But the taxes needed to do this will depress aggregate demand and so destroy other jobs.

But not necessarily the right ones. The jobs that need destroying are the bullshit jobs described by David Graeber; the producers of environmental, intellectual and risk pollution; DWP and Border Force staff who harrass benefit claimants and migrants; or those in the bloated financial sector. Job destruction, as I've argued, is not a task which can be left to the market.

It is, however, a question which Reeves is avoiding. Not only is she failing to see that "securonomics" and productivity growth are in tension, but she's not seeing that there must be losers from any serious attempts to raise economic growth - not just employers of low-wage workers, but also financiers and monopolists.

Yes, Reeves is breaking with New Labour in some respects, some of them welcome. But she's also continuing with them in at least one aspect - in not seeing that there are trade-offs and big conflicts of interest in economic policy-making.

Another thing: I'm also wary of Reeves' implicit assumption that we'll create lots of dynamic businesses if only we gave them easier access to finance. Whilst lack of finance is perhaps ones constraint on some firms, I suspect there are much bigger reasons for a lack of investment and entrepreneurship. But that's another story.