Becoming an SW1 NPC

Tom Forth, 26 May 2024.

I quite like Dominic Cummings. Not personally. I don’t know him. But I think we need more disruptive people in and around UK government and he has been one. He’s probably still got more disruption in him and I look forward to seeing what he does next.

I’m told he was annoying at best and a bully at worst when he was in government. That’s a shame, but it’s hard to contain disruption and even harder, I fear, to achieve much without overconfidence. Good teams include difficult people.

 

When Cummings published his famous misfits and weirdos blog post quite a few people told me I should apply. I was, they said, exactly what he should be looking for. My blog posts, especially the one on the heritability and importance of human intelligence to regional economic growth, are probably spicy enough to generate some calls for me to resign within a few weeks of appointment, though probably not controversial enough to force me to. So I might have worked out better than the people he actually picked.

I didn’t apply. In part because of how Dominic’s and my life have differed since their similar starts. My experiences make me reluctant to work in Westminster.

I’m quite a bit younger than Dominic, but we grew up not far from each other. If Wikipedia is to be trusted, our parents did remarkably similar jobs. We both got out of the Northern Eastern parts of England pretty much as soon as we could. Then things diverged.

From a private school, Cummings went to Oxford to study history.

From a state school, I went to Imperial College and Paris to study physics.

Perhaps we’d have both done well with a bit of each other’s experience. My impression is that he wishes he’d studied a bit of physics himself, though I wonder just how many weeks of matrix multiplication he’d have managed. I sometimes wish that I had the guilt-free confidence that private schooling and a non-technical degree at Oxford seems to bring to many of its products.

 

I went on to do a PhD in computational biology back in Yorkshire, working on much of what a decade later became called big data. I did quite a lot of work on what was called machine learning for a while and is now called AI. I’ve moved around since then. Leeds, Birmingham, Manchester, and now back to Leeds.

By the time I was being told I should talk to people about jobs in SW1 I’d built a career for myself here. Under the guidance of and in collaboration with Paul and Kathryn Connell, I was part of a team that built up The Open Data Institute Leeds to a national institution of some importance, though never with much national backing.

We’ve done so much stuff that I can only pick a few relevant things.

I wrote a report that helped make the case to boost public sector R&D funding in Britain. We built the first versions of an app for Teacher Tapp, a company that now surveys 10,000 teachers every day and is one of the most innovative and interesting startups in the country. We built some of the most-used apps in local government, such as Leeds Bins, and saved millions of pounds for local government in the process. We’ve tracked hundreds of millions of bus journeys in the West Midlands over seven years to improve bus services and helped make the case for the bus regulation which the Conservative government delivered with the Bus Services Act in 2017. We helped with the open data parts of that law that mean you can now track every bus in Britain in real time. And we helped the UK government release the LEO dataset which is now informing debates around the future of UK universities.

Now I am CTO, and own about a quarter of, The Data City, one of the UK’s fastest growing AI companies. We export data to companies and governments all around the world as well as sell into businesses and local and national governments in the UK. Our data enables exactly the kind of technology-powered government Dominic Cummings wants to build.

 

With that back story, I was amused, but also bemused, when Cummings decided to call me an SW1 NPC the other day. SW1 is a postcode district of London containing the UK’s Parliament, Number 10 Downing Street, and Whitehall: the centre of the government’s civil service. NPC refers to a non-playing character in video games and is used as an insult by the terminally online. It has various meanings.

The Economist has called me “one of Britain's most influential wonks”. That was probably a bit generous, but it’s not completely imagined. What I have achieved in that part of my life I have done very deliberately without spending time in or working in London.

It's not because I dislike the place. London is a great city and a centre of huge opportunity. But surrounded by most of the people anyone would need to talk to about almost anything in the world, I’ve found a tendency there to look inwards. And with one of the largest governments and institutional endowments in the world at its heart, I sometimes see people dampening their entrepreneurialism and taking the easier public subsidy. These tendencies always feel even stronger in Oxford and Cambridge — two of the UK’s most public sector dependent cities and the places where a high proportion of Britain’s people in SW1 went to University. Further North within the UK we have access to as many people as in London, though not as complete a set. We must look outward and abroad more often for our opportunities. And we do not have the largesse of the state to fall back on in the same way if we fail. So I have stayed away from SW1. If I am of that place nevertheless, it is not by any definition that I understand.

 

And if I am an NPC, then it is in significant part because instead of working my way into being a playing character I spent my time studying science, learning to write some of the best computer code in the world in my tiny niche of specialism, becoming an expert in AI and big data, starting and growing businesses, building things in both the private and the public sector, and gaining precisely the experience and skills that Cummings has often written are more needed in government.

Where I have worked with government it has overwhelmingly been with local governments. I find them more innovative, more agile, and much more efficient. Implementation and policy are not just closer together, they are often the same person. During Covid it took us a week to build and deploy an app that let us send instant messages for free to people on any and every street in Leeds. It cost £1000.

Working with local and devolved governments is much closer to working with a startup, though sadly one that is forbidden by the monopoly incumbent in their sector from raising more money. By comparison, working with national government feels like joining IBM, a stodgy monopoly resting on its former glories, absent the push of competition that drives innovation. It’s a valid choice — IBM have done and continue to do great work. But there is a reason why it is startups and not IBM who drive the fastest improvements and it is similarly why to do innovative things in Westminster you have to break the rules or “build a startup within government” before you can get to work.

Decentralisation

I came to Cummings’ attention recently because I responded to an interview he gave to The Independent newspaper. In it, and following his time working with focus groups around the country as he plans to set up a new national political party, he responded to a question,…

 

 

I completely agree. I have been writing about that and trying to fight the tide of domestic political centralisation in Britain for nearly twenty years. With a few exceptions, my side has been losing. And during that time one of the successful people on the other side has been Dominic Cummings.

He was part of the team that won a referendum in 2004 in which North East England, the region he grew up in, rejected a regional assembly. Power remained in Westminster.

He has done other things in his political career that I think have accelerated the centralisation of the British state, so I said so. He didn’t like that.

So I asked which decentralisation he was proudest of from his time in government. I wanted to know what things I was missing that he achieved during a time where he claims to have been focused on decentralising domestic politics, taking power away from government and organisations in London, and giving it back to government in Bolton or Birmingham. That’s the definition of decentralisation he uses in the interview.

You can read the full 1000 word reply online, but I’ll simplify and reply here for you if you’d rather not and if you trust me to fairly summarise it.

I asked for one, but Cummings claimed six major decentralisations of power which he is proudest of. None are decentralisations according to his definition.

1. Brexit.

Brexit is a decentralisation of government. It clearly removed power from Brussels and reclaimed it for Westminster, a lower tier of government. I voted to remain in the EU, but have never questioned the result, am glad we left once we voted to, and do not think we should consider rejoining for a generation. I wrote about the opportunities I saw in Brexit to decentralise the UK’s government and economy in May 2018.

But Brexit has not decentralised the government or economy of the UK. Domestic powers reclaimed from Brussels have not made their way to Bolton or Birmingham. European Regional Development Funds have not been fully replaced domestically, nor have extra powers to lower or raise local taxes been devolved to local government.

The pint-sized champagne bottles that Brussels banned were legalized by the UK government. But my governments in Leeds and West Yorkshire were not given the power to legalize a 600ml bottle of wine for consumption here or to ban those pint-sized bottles. There are very good reasons for that as I’ve written about in comparing decentralisation of standards in the USA and the UK. A single market based on centralised standards has long been a comparative advantage of the UK. But it is not decentralisation of power.

A larger frustration that I have with Brexit is that it seems to have led to such a large increase in the size of the civil service and its further centralisation in London.

2. Establishing free schools and 3. Protecting academy school autonomy from central government.

Both the Birmingham Corporation School Board and the Bolton Corporation School Board were set up in 1870. When that city and town were among the richest in the world they were providing most of their education. The removal of schools from local government control by the UK central government under both New Labour’s Academy system and the Conservative government’s Free school program may well have improved the quality of schooling. But it was not decentralisation by the definition Cummings uses in his interview.

The ONS agree with me and transfer school staff from local government to central government employment in our national statistics when a school transitions. This transfer explains a part of why since 2010 the UK’s central government has expanded so rapidly while local government has shrunk.

Some people argue that because free schools and academy schools remove control from local government and, via central government, give it to communities they are in fact a decentralisation. There is some merit to this argument, but it is neither the definition of decentralisation that Cummings uses in his interview, nor the definition of decentralisation used by international statistical organisations.

I find the argument against it both simple and conclusive. If education were decentralised to local government as in, say, the USA, those local governments could implement similar schemes at the local level. This indeed happens in the USA. I would certainly have argued that the EU government forcing Academy and Free schools on Britain would have been a centralisation of power. I think Cummings would agree. We would both have found it unacceptable.

4. Creating ARIA.

I don’t like writing reports. One of our missions at Open Innovations is to kill all reports. So when Richard Jones and I wrote The Missing £4bn for Nesta, still one of their most downloaded and read publications, I agreed to do so because the topic meant a lot to me.

I think that the UK gets much less return and impact from its fantastic scientific research than it could do. This is in part because we fund the wrong types of research, in the wrong places, via the wrong structures. Our report showed that, and it supported the creation of a new body like ARIA to try and get some money out of the research councils and university-focused research.

I know that Cummings read parts of the report, or at least got my co-Author, and briefly PhD supervisor, Richard Jones to come down to London to talk about it. I also know that our work helped win, though of course I’ll never know in what measure it helped, the argument to get ARIA created and then later funded, because I spoke to both James Phillips and senior civil servants about our work and did some extra analysis for Number 10 and for Cabinet meetings around that time.

I am glad that ARIA exists. So much so that I set up the Wikipedia page for it and after a number of invitations I wrote an application to be one of their founding PMs. I referenced decentralising power to Bolton's neighbour Wigan in that application and in my explanation of why I wouldn’t be applying.

When ARIA announced that it would be setting itself up in London, hiring lots of people in London, and getting its staff to meet in London, I was gutted. Gutted because there’s so much data that shows that public sector R&D allocation is biased by location, and gutted because we know that London already receives far too much public sector R&D funding given its low private sector R&D spending. I still used one of the four work trips to London per year I allow myself to attend their opening to try and convince myself that the location choice was okay. I couldn't.

ARIA is not a decentralisation of power or money to Bolton or Birmingham. It is a reorganisation of money that was controlled in London and remains controlled in London today. I still think it’s a net positive, but it’s not decentralisation.

5. Deploying rapid testing and centralised vaccine rollout during Covid.

The UK government’s record on Covid is bad. By our best guess, more people died than in any of our neighbours. And by our best guess, more people died in England than in Scotland, Wales, or Northern Ireland which significantly opted out of the UK’s centralised Covid response.

Paper after paper, article after article, and report after report blame part of England’s poor performance on the extreme centralisation of its government response. Local public health teams were cut out from the pandemic response and replaced with inexperienced centralised services that didn’t work. Centralised data services failed at critical moments and tight central control of messaging left people unable to respond, share local information, or trust government at the time it was most important they did. It remains a sign of great national decline that so much of the SW1 blob celebrates a functional central government data dasbhboard while ignoring its correlation with our high death rate.

Not everything about the UK’s incredibly centralised systems is bad. The vaccine program worked well for example. There’s no way that a decentralised vaccine development and procurement system could have worked as well. I wrote a blog post at the time about how despite centralisation costing a lot of English lives, it may well have saved far more lives around the world.

It’s great if Cummings helped fix some of the worst parts of the centralised system, like postings out Covid tests for people to input the results of into their national app. But again, it’s not decentralisation as defined in the interview. This isn’t just nitpicking or semantic games. Improving a centralised system is not decentralisation.

6. A new management system for Number 10, 7. Prioritising knowledge over seniority in Cabinet meetings, and 8. Opening up the civil service to talent and restoring individual responsibility.

This all sounds like good management. Lots of local governments are doing this and much more, in competition and under severe financial pressure. It’s great if the UK central government is also doing it with a monopoly on power and during rapid expansion. But it’s not decentralisation of domestic power to Bolton or Birmingham.

There was another point about the work done on reforming planning in England to make it easier to build things. But the policy got cancelled so why include it on a list of proudest achievements?

 

A legacy of centralisation and decline.

Words matter. Precision matters. Focus matters. Being right matters.

In more and more of his recent writings and interviews Cummings has spoken about the decline of these traits in British central government. I agree with him. But his replies to me exhibit the same faults. They are the type of wordcel bluffing that he complains about.

Deliberately getting my name wrong, and not correcting it when told so, was rude. But more importantly it showed a lack of care about precision. A lack of focus. A disinterest in being right.

Most importantly, none of the decentralisations he is proudest of match the definition of decentralisation that he defines, correctly, in his interview. They are attempts to improve the centre of central government in London, not get power back out of London to local governments in places like Bolton and Birmingham.

This is a shame because I think that the UK government has a decent story to tell on decentralisation. Welsh devolution was strengthened via a successful referendum in 2011. George Osborne forced through the creation of Metro Mayors and devolution deals for city regions in England in 2014. These deals have continued to be signed and now cover at least 60% of the population of England. Theresa May delivered on the promised return of bus franchising powers to city region governments in 2017. Two of these three achievements would have been unlikely under Labour.

And the record of decentralisation when it happens in Britain is good.

Since voting for devolution, Scotland’s economy has outperformed England’s, outgrowing every region except London and overtaking both South West England and East England to become the third strongest regional economy of Britain.

Since voting against devolution, North East England’s economy has grown the slowest of any British region and it has fallen behind Wales to become the weakest regional economy in Britain. The area has never been more reliant on money sent from London or more controlled by decisions taken there.

A real start-up party

My guess is that in preparing to launch his new Startup Party, Cummings has been running a lot of focus groups. And he’s heard what British people, and English people especially, have been saying for decades. That they’re sick of their lives being run from Westminster, by people they don’t know, who aren’t invested in the success of the rest of the country, and who consistently make bad decisions that keep them poorer. Now he’s convincing himself that he’s always known that, telling himself that he’s always worked to decentralise government, and getting ready to tell them that he’s on their side.

It’s delusional, but it’s got a chance of working. He's won two referendums already. He might get lucky a third time.

I hope he takes a different opportunity though. Get out of London, break out of the SW1 bubble, have a few pints with me and a few other NPCs working in AI and with local government in Leeds. Get my name right ideally.

It’ll be a chance to see that the real startup opportunity in government isn’t in Westminster, it’s with the tier of government that faces more of the same competitive pressures, the same resource constraints, and the same burning platform incentives to innovate that startups do.

A real start-up party would start in local government and work its way up. Wouldn’t it?