Andy Haldane worked at the Bank of England from 1989 to 2021, rising to the position of Chief Economist. He is now Chair of the Royal Society of Arts, and – subsequent to this interview – served as Head of the Levelling Up Taskforce for the UK Government.
This interview was conducted on 26 July 2021.
Q: Can you tell us about your role in policy over the past few decades?
My role has been principally with the Bank of England, for the last seven years as Chief Economist there. Prior to that I was in charge of the financial stability area at the Bank, with for setting monetary policy, and prior to that for designing and setting financial policy.
Growth, as you know, is a secondary objective of both the Financial Policy Committee and the Monetary Policy Committee (‘MPC’). When it came to all matters MPC, the growth picture – albeit the aggregate growth picture – was a key input.
The other way in which I’ve directly had line of sight on these issues was when I chaired the government’s Industrial Strategy Council, which was set up in 2018 as a way of monitoring and tracking process on the government’s Industrial Strategy of 2017. That Council – indeed, the Industrial Strategy itself – was abolished earlier this year. But in that role, for a couple of years, we had a lot of research looking at the patterns of growth, both aggregate and regional and sectoral. It was looking at the drivers of success and indeed failure, both historically in the UK and internationally. We did a lot of international case study work to look at, for example, how ‘what-they-didn’t-call levelling up’ was happening in various places around the world.
Q: when you look over that period, what do you conclude about what worked and what didn’t?
The three common ingredients for success or failure, looking at experience not just from the UK but internationally, is having regional policies that are sufficient in scale to the challenge at hand. That is not to be sniffed at, because the natural forces of agglomeration would, without a countervailing strong hand from regional policy, lead to rising inequalities, regionally especially. And that’s what we’ve seen. The powers of agglomeration, both economies of scale and scope, have probably become more potent and powerful over time, as super cities have benefited from a kind of global tailwind.
That has meant that regional policy would have had to have run fast just for regional equality to stand still. In practice that has that has not been the case along three dimensions. One is scale. The scale of resources set aside for this has not met the challenge. Second, duration. What we’ve seen in the UK, not just since 1979, but in fact since 1945, is a chopping and changing of regional policy. This almost preordains failure because you’re talking about structural policy here, and structural policy takes not just years but in some cases decades to take root and reverse those forces, and we haven’t stuck with anything long enough for it to fully bear fruit.
The scale and longevity points in some ways are a statement of the obvious. The UK is a notable outlier relative to other large, advanced economies in how much chopping and changing there has been. I forget the Institute for Government’s estimate of how many stripes and flavours of regional industrial policy we’ve had, but it runs into double digits since the Second World War – in other words, far too many.
The third common ingredient, which is a bit less focused on but which I think is every bit if not more important, is the capacity to join the policy dots. If there were a single secret sauce for turning the corner on growth regionally or nationally, we’d have long since discovered it. If redemption lay in, ‘let’s build loads of hard infrastructure’, then we’d have long since solved that problem. But the lesson from other countries is very much that you need a range of policy arms to be acting in a coordinated fashion to break into the problem of regional growth or levelling up or whatever you might call it.
Of course, that includes hard infrastructure. These days increasingly it includes digital infrastructure. It absolutely includes what these days often goes by the name of social infrastructure, which means having places people might want to live, so green spaces, decent high streets, parks, museums, community centres, youth clubs, football clubs, etc. That social infrastructure is increasingly being seen as a core ingredient of making a success of this. Skills and business, that goes without saying. And if any of those ingredients are missing, I think other countries internationally have found it devilishly difficult to reverse things, actually. So that’s a sense in which the ‘joined-upness’ of policy is a third crucial common ingredient. And we also know that, if that’s done nationally, it’s desperately difficult to do because Whitehall departments find it jolly tricky to do something joined up that would call itself a strategy as distinct from a set of policies.
That’s a very long-winded answer to your question but scale, longevity, interdependence of policy, appear to be the core ingredients of both success and failure. Because UK PLC’s done a, in relative terms, poor job of all three, it’s done a relatively poor job in terms of regional inequalities.
Q: How much convergence across regions or places do you think is achievable? And how much do you think is desirable?
I think picking a level on this is tricky. There’s nothing intrinsic about UK PLC that should lead us to expect so much wider regional differences than some of our competitors around the world, nor why those measures of inequality have been rising for the last half century, and are back, basically, to late Victorian levels. I don’t know what level to ‘level up’ to, but I do know I wouldn’t want those differences to be rising over time. I’d want them to be falling over time towards levels that were more broadly comparable to some of our, as I say, international competitors. I know that’s not a quantitative answer to your question, but I think the direction of travel matters almost as much as the destination when it comes to these questions.
Q: How much of this reflects the particular role on London in the UK’s economy, unlike peer countries except perhaps France with Paris?
Certainly, in an arithmetic sense I think there is some force to the structural characteristics of the UK. We have a first city that’s so, so much larger than our selection of second cities. Of course, there’s a fight on who even calls themselves the second city in the UK. But whoever it is, they are a tenth the size of the first city. That much is plainly true, at least in population terms. And in fact, most of those in contention to be 2nd, 3rd, 4th and 5th cities across the UK have, if anything, punched below their weight. That is to say, their performance on measures of pay or productivity or income, would be less than you might otherwise expect from a city of their size.
If you draw that international line of best fit of, say, productivity versus population size, what’s notable is how many of the UK’s bigger, and on the face of it, well-performing cities – I mean Birmingham, I mean Manchester, I mean Leeds, I mean Bristol – lie below that international line of best fit. I think that’s quite interesting when it comes to thinking about what is going on.
One story might be that London, although you wouldn’t sometimes believe it, has sorted out its local transport networks. Whereas local transport networks, among our 2nd, 3rd, 4th, are basically pathetic. They haven’t been invested in in anything like the same degree, Birmingham being a case in point. Greater Birmingham has working population that’s pushing up towards two million people. But because it’s weak local transport networks, its effective working population means it operates more like a one-million-person city than a two-million-person city. That does speak to the importance of local transport networks. Often, they matter very much more to local regional productivity than do the signature big ticket infrastructure projects that we hear about which costs so much up and down the country.
The point I really want to make though, about ‘London versus the rest’, is that, on the basis of the work I’ve done, I think sometimes these special stereotypes are neither accurate nor helpful. If you cut the cake as London vs the rest, you perhaps get that answer. If you cut the cake, ‘city versus town’, you don’t really so much. And indeed, if you look even within the cities performing well, the patterns of spatial inequality are often larger within cities than they are between city and town or than they are between regions. So, for me, the map of the UK is an incredibly variegated one where even within, on the face of it, well performing cities, you have pockets of tremendous affluence sitting cheek by jowl with areas still of multiple deprivation. That’s less true in London now than it was 10, 20 years ago, though still true to a degree, and it’s certainly true of all the cities I mentioned earlier on which, on the face of it, you’d say were performing well. So, when I’ve thought about this, I must admit I’ve gravitated away from any strict ‘London/non-London’, ‘North/South’, ‘city/town/rural’ divide. I think that the map is rather more complex than that, and therefore I think the pattern of intervention would need to be rather more complex than that to kind of break into these inequality problems. Hyperlocal is what some people call it, I think there is a hyperlocality across the UK.
Q: What implications for governance do you think follow from this variegation?
It strikes me that if we have got this hyperlocal economy – which we have, it’s almost fractal actually, because the more you hold your magnifying glass up to it, the more you see the same patterns self-replicating at a very local level – then there’s no way any centralised policy apparatus has any hope of addressing those increasingly hyper local problems.
In the same way as the regional disparity pattern has been widening for the last half century, the same has been true at the sub-regional level as well. That pattern is replicated, in fractal fashion, there too. If you look at other countries, as I have, or other regions that have made a success of closing those regional disparities, that has not been possible without ‘devo-max’ relative to anything that we are close to right now. Only those local areas have the local knowledge necessary to tackle hyperlocal problems. Only those local governance agencies, in their many and various forms, have the agency as well as the information to deliver change on that front.
A third dimension which speaks to my point about interdependency is it actually is much easier to join the dots, join the different arms of policy, when it’s operating at a regional level or a local level rather than national level. Actually, they find the coordination of policy arms that much easier than is possible from a centralised infrastructure. The point you have heard a million times, which was repeated by the prime minister last week, is that it is no coincidence the UK distinguishes itself both in its degree of centralisation of policy and its degree of regional inequality. Those are not coincidences for me. A much more root and branch devolution of the holy trinity of monies, powers and peoples is an absolute sine qua non of making any inroads into levelling up. All the experience I’ve seen internationally speaks to that. It’s not the only ingredient. It’s certainly not the only thing you need to do. I think without it you’ll have a devil of a job tackling the problems we’ve been talking about.
Q: Have there been periods of time when Whitehall has understood, and risen to, this challenge?
I see the occasional flicker. I don’t know that’s necessarily been through lack of interest. The Board of Trade under [Michael] Heseltine had aspirations on this front. Vince Cable in the Coalition had some aspirations on this front. Most recently, under Theresa May had, Greg Clark had aspirations on this front.
I think in each and every case, they probably fell foul of not satisfying those three criteria mentioned earlier on. None were given enough financial room to achieve the scale. None were around in office for long enough to see the job through, even if they were doing things. Take the 2017 Industrial Strategy. Greg had intended that to have been a way of joining up Whitehall, but the rest of Whitehall wasn’t ready for being joined up. So, it didn’t happen. And all that happened on that front was the stuff that the Department of Business, Energy and Industrial Strategy (‘BEIS’) and only BEIS could do – on a scale that was subscale, because they didn’t hold the chequebook. And of course, the strategy itself didn’t have legs. It was gone within three years.
Q: Is this cross-departmental way of working feasible?
It is a big ask. It goes against the grain of much within the Civil Service. It would need – necessarily, I think – to be led from the very top for that to happen, and to have the Treasury’s willingness with its own chequebook to make good on that. I just don’t know that the stars have fully aligned on that at any time I can think of in the post-war period. There may have been moments where it’s been true, but you need to sustain that through a few cycles for it ever to get rooted in Whitehall psychology and I don’t think – for those reasons – that it has.
As I say, I do think, when I go to Manchester – not suggesting Manchester is by any means perfect – that they are having a joined-up conversation about health, about education, about transport, about higher education and further education, about the role of civil society, about business. It was really striking the differences in quality of Local Industrial Strategies, ranging right through from Manchester – top of the pile, very well developed, very joined up – to a very long tail of, in this case, Local Enterprise Partnerships (‘LEPs’) or mayoral combined authorities, of industrial strategies which were really quite thin. That told me two things: one, the joining-up was possible, because Manchester and West Midlands, to some extent one or two others, are making it happen. They’re having that conversation, not perfect but better than you might otherwise get through a Whitehall-led effort. But two, that capability slides away very rapidly beyond probably Manchester and the West Midlands, maybe one or two more. That leaves large swathes of the country without the means to draw up plans.
Q: Is the metro mayoral model right here? Is the scale right? What about the places ‘left out’ of administrative boundaries?
That is a great question to which I plainly do not have a great answer, so I’ll not pretend that I do. I do think that the mayoral model is something that can achieve scale, critical mass. A rolling out of something at around that level to the places that don’t currently have a metro mayor – I think more than half the country is not covered by a metro mayor currently – was actually mentioned in the Prime Minister’s speech last week as an aspiration.
I’m thinking something of that size is needed initially to generate the capacity. The reason we don’t have mayoral combined authorities in the other two-thirds of England is because that critical mass is not there right now. Some of my advice to government has been the same advice that the World Bank learned 50 years ago: you need to put money into building capacity and capability on the ground before you decide on distributing monies and powers.
Outside the places that already have mayors and are thriving, I would be making up-front investment in the ‘people’ part of that Holy Trinity – before you start delegating the monies and the powers – because I think without that it’s really tough for them to form a plan that makes sense and to execute on that plan in a way that makes sense.
I think once you have got that – you’ve filled the map with mayoralties of a critical mass, with the peoples, with the powers, with the monies – it’s absolutely possible to start having a conversation about delegating down to a further level within those geographies, and therefore addressing some of those hyperlocal problems at a hyperlocal level. I wouldn’t go immediately to a map of the country where decision making was at the same level of problems. I think that would make for a jigsaw puzzle that wouldn’t fit together.
Q: How much should we think of cities as the drivers of their regional economy, as opposed to a more balanced approach across a given region?
I think we absolutely should want to harvest the benefits of agglomeration. In other words, we’d want our cities to be as well-performing as they could be. There’s work to do on that front, given the way the cities currently punch below their weight in international terms. I think that for any strategy for levelling up, supporting growth of our cities ought to be a key ingredient. We wouldn’t want to level up by levelling them down. That would be quite counterproductive. There’s lots of scope to harvest some more growth gains from, in particular, non-London, larger cities across the UK.
Outside of that, pre-COVID I’d tell myself a story about how, with sufficiently effective local transport routes, city-regions could spread their tentacles really quite far and that would generate, as we’ve seen across some places, widely-spread spillover benefits to towns, and perhaps even some local cities that might otherwise be struggling. I think that is a viable strategy from some areas. You could basically broaden the circumference of what counts as a city-region and therefore get many of those spillover benefits, that the agglomeration benefits of cities spillover to a wider cohort.
In the light of COVID, maybe we can rethink that. What it demonstrated is that physical distance, at least five days a week, is less essential than we might think, and therefore some of the suburban or even rural retreats might be viable within an even wider spread of that city-region, than was otherwise the case. I think that’s worth thinking about. (Not that working from home and digital infrastructure will entirely reverse the forces of agglomeration, but they at least may have weakened them somewhat when it comes to the location of both business and workers in those businesses.)
Even with that, there will be the case that some places are still out of scope, far from being easily able to attach themselves to any nearby city. The South West will be a case in point, which doesn’t really have a city anywhere close that would count as a hub. If you speak to people in Cornwall and Devon, they’ll tell you this. There are ways of having connectivity without it being this hub-and-spoke type model, and that’s what they are seeking to develop. If the common factor is not an agglomeration of activity in the city, but the common factor is an identity around what it is that they produce, then having a strategy that’s not city-centric but is centred on a given sector, or given industry, or given activity, possibly carried out in a quite decentralised way, strikes me as a completely viable strategy for some parts of the country that don’t have a city plonked in the middle of them.
I think there are a variety of possible models there. The key common factor, though, is that everywhere needs something different. For some, it might be a city centre that’s thriving. For others, it might be a common identity or sector or specialism, that is distinctive to them. If I were forming an industrial strategy right now I would want to make sure that each of my metro mayors – that now suddenly fill the map – have something that’s distinctively theirs, by way of specialism, or sector, or industry, and want to invest in that regional specialism. Not just because it creates jobs, but as importantly because it creates identity for that region away that I think lots of regions currently lack. That’s what other countries do as well, by the way. That’s unsurprising, nothing I say today is actually my idea. This is basically what other countries have been doing for quite some while.
Q: What are your conclusions from success stories in other countries?
The Industrial Strategy Council did a nice report on this. There’s always a risk that you cherry-pick things that have worked and don’t focus enough on things that haven’t and there’s many more things that haven’t worked regionally than there are that have. We did six case studies from around the world. The German case study I quite like.
One that we don’t cover there, but which I like in some ways even more, and I know you can’t read across exactly on this, is the Singaporean experience with their approach to industrial policy. We can’t replicate Singapore for all sorts of reasons, some of them good and some of them less good. Nonetheless, it has a proper approach to strategy: up front, I set myself a mission, and my mission is to be a world-leading producer of X, where X is picked in a very enlightened way. You then set about, as an act of strategy, building the ecosystem, not just the industry, but the supporting supply chain for that industry, over time. You actively make a point of stealing the best – stealing, begging, borrowing ideas from elsewhere around the world as a way of informing where the frontier for that sector is, and as a way of forging commercial links internationally with that sector. And then you pursue it – at scale – for a five- or eight-year period. It has scale, it has longevity, it has that coordination I mentioned earlier on. It has the overt backing of government, which helps no end as well. If there was something I’d be seeking to come closest to replicating here in the UK, it will be something like the Singaporean model.
Q: On the German case study, how applicable is this to the UK given the longer history of decentralisation in Germany, and the lack of a dominant city?
They do have those natural governance/political advantages. They’d be hard for us to replicate lock, stock and barrel, but could we do something that moves in that direction? That strikes me as eminently feasible, without seeking to move to their federal model in quite that way.
It’s not just the governance, but the supporting anchor institutions that I think matter every bit as much. That includes anchor institutions for financing of industry and financing of jobs. So their three layer cake, of the Sparkasseländes Bank, and KfW Development Bank. They have the sub-regional, regional and national, but decentralised. They have that, we have nothing, basically.
They have their Fraunhofer Institute, of which there are many more than there are Catapults here in the UK, and they’re of a scale that’s an order of magnitude larger. And they are there to facilitate both innovation in the same way as our Catapults, but as importantly, ensuring that innovation trickles down to the small mass of smaller businesses. Right now, our innovation engines, which includes the Catapults, and will probably include the new innovation agency as well, very much service the needs of bigger businesses already at, or close to, the frontier. The Fraunhofer do as much of a job in making sure the fruits of innovation trickle down to the mass of German SMEs.
We have just got that wrong. We have not put in the institutional infrastructure necessary to support technological trickle down. That’s another reason I think Germany has done a better job of smoothing within-region and between-region inequalities.
My stylised way of thinking about it is that you can think about the bifurcations in Britain along three axes. There’s one operating at the corporate level between a bunch of high-flying businesses with high-flying productivity on the high road, and then a long tail of languishing companies in productivity terms. The gap between those has been widening over time and that’s the main reason why productivity is flatlined in the UK.
Then we have the same bifurcation operating at the level of workers within those companies. We have those on the skills high-road – let’s call it the half of young people that go to university – and the half on the skills low-road who don’t go to university.
Then the third axis, or third point of the triangle, is the spatial dimension between high-flying regions, or high-flying cities, and less well-performing places around the UK.
The reason I’m mentioning this model is because those are three points of the same triangle. The high-flying companies tend to employ the workers on these skills high road, and tend to locate themselves in the better performing parts of the UK. The flip is true of the flat-lining, long-tail set of companies who tend to employ workers on the skills low-road and locate themselves in less well performing parts of the UK.
I say all of that just to make clear that the spatial dimension of this has as its mirror image what’s happening on the productivity side among companies and on the skill side among workers. Unless we find ways of closing the gap on the on the productivity side among companies, which is why I mentioned the diffusion engines of the Fraunhofer, you’ve got no way of making inroads into the spatial differences either. The most effective ways I’ve seen internationally of closing those skills and productivity gaps is by having within each locality an anchor institution whose job it is to make that happen. And those things we currently lack as a country.
Q: Does it follow that if we make huge progress on skills, without building these local clusters, we will inevitably see a brain drain?
I think so. We know how much more footloose the graduate end of the labour market is. When you look at measures of labour mobility, look at the distribution of skills across the workforce and how the movement takes place, again the UK distinguishes itself from, say, Germany. In Germany, it is the case that a highly skilled, high-flying, youngish person might move from one of the big German companies, gilt-edge companies, to run a middle sized Mittelstand company. That’s the way they cut their teeth and make their name. You get this much greater mobility up and down the size and skill spectrum in the labour market in Germany than you do in the UK. The flow matrix in the UK jobs market is very telling, and basically the same high calibre graduates float around the same hyperactivity companies. And the same is true at the bottom end of the distribution and we need to find more cross diagonals in that transition matrix, both spatially and skill wise.
Q: What’s your overall assessment of regional growth policy over the last 40 years? Is there a particular idea or innovation that you think has really worked?
There’s been too little regional policy, and the volatility around it has been too large. The mean has been too low and the standard deviation has been too large when it comes to UK regional policy. Therefore, we need to both raise the mean and lower the volatility to put us on a more even keel. There’s lots of things that can and need to happen to change that.
If you’re offering me one wish, I think you’d have to start with what we’ve discussed, namely a devolution agenda in its many and various shapes and forms. I don’t like the language of ‘doing deals’. I think that relationship between central government and local government – and I see this this now in the various funding pots out there, such as the Levelling Up Fund and Town Fund, i.e., centralised pots competitively bid – isn’t the way to do it.
It doesn’t work, first because you’re then finding the monies go to those best able to bid, which is the opposite of levelling up. It’s back to the capacity and capability question. Second, you want this relationship not to be contractual. You’re not ‘doing a deal’. This is more like a covenant between different parts of the country. You can’t run this like an outsourced business, giving tenders for your services. That’s just feels to me like the wrong relationship, tone, to be striking. You need a relationship of trust: we are empowering you, giving you agency over decisions and over monies. And you will discharge them in a way that best meets local needs. That feels more right to me.
Mayors may be the nth best solution, but where we are now is n minus one or two or 10 or whatever it might be. Fill the map. I worry right now that we’ve got a different sort of inequality, which is that those places with mayors can get their act together, and those without can’t. That’s why a more comprehensive, consistent map is important.
ENDS