In the debate about how to increase productivity, everybody seems to be making an assumption that I find questionable - that feasible policy changes can make a significant difference.

We need some historical context here. My chart shows that big, sustained productivity growth is in fact very rare; growth of over 2.5 per cent a year was only seen over lengthy periods between 1945 and 1990. Our idea that rapid growth is normal and stagnation not might owe more to the influence of our formative years than it does to the wider historical evidence.

To see why strong growth should be so unusual, imagine an industry could treble its productivity over ten years. If it accounted for two per cent of the economy, it would add less than 0.3 percentage points to total annual productivity growth. This tells us that a few dynamic sectors are not enough to achieve good aggregate growth. Productivity growth was minuscule in the early decades of the industrial revolution because fast-growing sectors were only a small part of the economy.

Which is the problem we face now, in modified form. The economy is dominated by services, in many of which efficiency gains are terribly difficult to achieve: teachers, waiters, care workers and hairdressers have few ways of significantly improving productivity: this is Baumol's cost disease. As Dietrich Vollrath writes:

Most service industries have relatively low productivity growth, and most goods-producing industries have relatively high productivity growth. As we shifted our spending from goods to services then, this pulled down overall productivity growth. (Fully Grown, p5)

Perhaps, then, strong productivity growth was the aberration and stagnation is the norm. Certainly, this is what classical economists thought. All of them foresaw a "stationary state" in which growth ceased because the forces of diminishing returns outweighed technical progress. As John Stuart Mill wrote:

It must always have been seen, more or less distinctly, by political economists, that the increase of wealth is not boundless: that at the end of what they term the progressive state lies the stationary state, that all progress in wealth is but a postponement of this, and that each step in advance is an approach to it.

Marx's view that a falling rate of profit would ultimately lead to capitalist stagnation was essentially an elaboration of his predecessors' thinking: Marx's admirers and detractors both under-appreciate the kernel of truth in Paul Samuelson's gibe that he was a "minor post-Ricardian."

It's not just Baumol's disease and diminishing returns that limit productivity improvements, however. There's also a collective action problem. Companies will only invest to raise profits if they believe (rationally or not) that there's a profit for themselves in it. This obviously limits investments that have spillovers: as William Nordhaus showed, firms capture only a "miniscule fraction" of the total returns to innovation. It can also curb investment if firms fear that future change will render today's technologies obsolete: why spend £1m on a robot today if it will be competing against a £500,000 one in a few years' time? And then there's the likelihood that a series of crises - the tech crash, financial crisis and pandemic - have depressed animal spirits.

All this means that even if productivity-enhancing technologies are available, companies have good reason not to use them.

There's another obstacle to productivity growth. As thinkers such as Mancur Olson and Joseph Schumpeter pointed out, economic growth produces powerful groups with vested interests in blocking further progress. Joel Mokyr calls this Cardwell's law:

Technological progress encounters resistance from various groups that believe they stand to lose from innovation. These pressure groups will try to manipulate the political system to suppress successful innovation....Under fairly general conditions, it can be shown that the single economy will move inexorably to an absorbing barrier of technological stagnation.

So, for example, big companies push for harsh intellectual property laws to preserve their monopoly profits, at the expense of expanding knowledge: there's a reason why that link to Mokyr's paper takes you to Scihub. Bosses prefer to use new technologies for surveillance and guard labour (pdf) rather than for expansionary or liberationary purposes. And bankers want high-powered incentives that boost their incomes even if these come at the risk of financial crises that contribute to years of stagnation.

Such conservatism is especially restrictive in our rentier-dominated economy. Rentiers oppose some changes that might raise productivity, such as shifting taxes from incomes to land. They also want low interest rates rather than the high aggregate demand that might drive productivity gains. In the post-war years General Motors needed a large well-paid working class; the latter-day Goldman Sachs does not. All of which blocks productivity improvements because it is this minority of people who have disproportionate political power.

There's a reason why Johnson's attempt to raise productivity by restricting immigration is so daft: more intelligent ways of doing so are countermanded by our political system.

All of which suggests that Marx might have been right:

At a certain stage of development, the material productive forces of society come into conflict with the existing relations of production or - this merely expresses the same thing in legal terms - with the property relations within the framework of which they have operated hitherto. From forms of development of the productive forces these relations turn into their fetters.

Capitalism was once a progressive force, but - especially perhaps in the UK - it no longer is. And this is due not to external shocks but to endogenous developments.

The point of this is of course not to say that if we had a revolution we could all live happily ever after. It is instead to pose a challenge to anybody discussing the UK's productivity problem. It's not sufficient to ask what measures might raise productivity: there are countless good candidates. We must also ask: are those policies possible within the constraints of the UK's political system? Technocrats must take class politics more seriously.