Global Economic Outlook Panel – panel remarks by Andrew Bailey

Speech

This year has marked forty years since I first joined the Bank of England. Before that, I was an economic historian. I am going to draw on some of that past in my remarks today.

In recent years the nature of the economic shocks we face has changed. The natural territory of monetary policy is to respond to shocks to aggregate demand and its components, set against a backdrop of a relatively stable and slow-moving supply side of the economy. That picture has changed, and we have seen and are seeing, large and faster moving supply shocks. We don’t have to look far for examples – Covid, Russia’s illegal invasion of Ukraine, tariff increases etc^{footnote [1]}. Such events prompt challenging questions on economic consequences. What impact has Covid had on longer-run labour supply? What impact are tariffs having on price setting?

There are also deeper supply shocks at work affecting the structural parameters of economies. Two are important to highlight. First, the negative labour supply impact of long-run population ageing. Second, the lower rate of productivity growth in many economies, the UK included, over the last fifteen years. In the UK, over that period the potential growth rate of the economy has declined from around 2½% per annum in the preceding twenty years to around 1½% since then. The largest contribution to that decline has come from productivity growth. It’s important to understand the impact of this change. For monetary policy it affects the speed limit at which the economy can operate. Of course, the challenge for much of the time has not been to restrain demand growth to that lower speed limit but rather the opposite.

Let me add another illustration here, using the UK economy. This is an admittedly naïve counterfactual, I make no pretence of sophistication. Imagine that throughout the last fifteen years the economy had run at a 2½% speed limit throughout (i.e. supply and demand had both grown at 2½% p.a.). Holding public borrowing fixed, the debt to GDP ratio would now be 82% instead of 96%, and the official projection for 2029-30 would be 79% not 96%. It’s a very simple point really. If the denominator grows more slowly, economic policymaking gets more difficult.

I’m going to revert to being an economic historian for a short while. Why? Well, because of four big lessons that we can take from history.

First, economic historians would be surprised at the lesser attention paid to the supply side, for the reason that over time longer-run growth has come from the supply side. There was a long debate on whether it was more a story of supply or demand in the British Industrial Revolution, and the answer was pretty clearly resolved – supply did it^{footnote [2]}. So we should pay attention to the supply side.

Second, supply side growth has not been sustained over time. It comes in waves, with the real dial-moving changes being caused by so-called General Purpose Technology innovations – starting with the steam engine, moving through electricity, and eventually on to the internet. Moreover, it hasn’t come quickly, though there are good reasons to think the speed of delivery is increasing. It is reckoned to have taken 40 years from Edison first wiring up a light bulb to electricity showing up in the productivity statistics.

Edison himself is supposed to have said over the length of time it took him: “I have not failed. I’ve just found 10,000 ways that won’t work”. We are still at the experimentation stage with AI, so investment and persistence is crucial.

The third lesson from history is that supply side growth across sectors of the economy is typically very uneven. Another of the big debates on the British Industrial Revolution was whether and why aggregate growth was slower than had been thought. It was, but not in the leading sector of cotton textiles^{footnote [3]}. An economic historian would not therefore be surprised by recent evidence from the US economy of slow overall growth in manufacturing productivity, a concentration of that growth in a few computer related industries which are growing rapidly notwithstanding the lower overall average, and methodological disagreements on how to measure that productivity growth (it may be even faster in the leading sector)^{footnote [4]}. An economic historian might now be saying something about déjà vu.

The fourth lesson from history seeks to draw out different growth models, and what we learn for today. The Yale economic historian Bill Parker used to teach economic growth in European history around three theorists of the past: Thomas Malthus, Adam Smith and Joseph Schumpeter. I’m going to finish by seeking to explain why this is more important than ever in reading today’s world economy.

Briefly, Malthus symbolised the theory that population growth had the potential to outrun food supply and thus limit economic growth. He was wrong, because productivity in agriculture increased and freed productive capacity to enable the Industrial Revolution. But I will come back to interpreting his relevance today.

Smith represented the idea that the extent of the market limits the division of labour, with higher productivity coming from economies of scale and specialisation – the famous division of labour in the pin factory. The key element of Smith’s thinking here was that a larger trade area would lead to more productivity growth and thus a higher potential growth rate.

But Smith and the other classical economists did not incorporate into their thinking the last leg of the argument, namely that continuous though fluctuating technological change and innovation, financed by the extension of credit would raise productivity and potential growth – in other words the Industrial Revolution and the following waves of General Purpose Technology. This leg of the argument is often associated first with Schumpeter^{footnote [5]}.

These lessons of history are I think highly relevant to the situation we find ourselves in today in respect of the supply side. Let me finish by briefly drawing out the key points.

First, we face universally populations that are on average ageing. It’s Malthus in reverse. Rather than the labour supply outstripping other essential elements of the supply side (food supply), we face it declining in a way that reduces the support for potential growth – thus putting more stress on the 2½% to 1½% issue.

Second, tariffs contradict the Smithian growth model, as do other measures that restrict trade. I can give a personal experience of this issue at work. For nearly a decade I have been very careful to say that I take no position per se on Brexit, which was a decision by the people of the UK, and it is our job as public officials to implement it.

But, if you ask me what the impact is on economic growth, I do have to answer that question as a public official. And the answer is that for the foreseeable future it is negative, but over longer time there should be a positive, albeit partial, counterbalance. Why? Because the growth model of Adam Smith is clear on this point. Make an economy less open and it will restrict growth, though over a longer time trade will adjust and rebuild. And, this appears to be what has happened. The same argument holds for the world economy and tariffs.

Third, and back to the 2½% to 1½% point, if we take into account the impact of ageing and trade restrictions, we are putting our chips down on the Schumpeterian growth model. So, we need to foster investment in our economies. And, since AI looks like it may well be the next General Purpose Technology, we must work with it and ensure that it develops appropriately and well. It is timely, therefore, that this year’s Nobel Prize in Economics was awarded to Joel Mokyr – whose work on the prerequisites for lasting growth has demonstrated the importance of scientific discovery interacting with practical knowledge in a society that welcomes technological change (as the Royal Swedish Academy of Sciences described it) – and to Phillippe Aghion and Peter Howitt for showing how technology progresses through the innovative dynamics between competing firms.^{footnote [6]} On this point, I will finish by adding that there is nothing inconsistent with thinking that AI is the next big technology, and being concerned that it may along the way challenge financial stability through stretched valuations, and particularly in an environment of larger global supply shocks^{footnote [7]}.

Thank you.

I would like to thank Sarah Breeden, Fabrizio Cadamagnani, Karen Jude and Martin Seneca for their comments and help in the preparation of these remarks.