AI risks

This, from Tyler Cowen, has been doing the rounds: he tells us not to worry about artificial intelligence. Scott Alexander takes issue with it here, Noah Smith broadly agrees with it here (albeit, in typical Smith fashion, he disagrees with the most obviously correct bit of it), Zvi disagrees at really quite some length here and then Tyler replies to Scott here. No doubt the usual suspects have been chipping in elsewhere as well. So I don't feel too bad about adding my two bytes. 

In short, I'm not worried about AI. Perhaps I should be, but I shall try to explain why I'm not.

Why are small countries rich?

In my post about Boosters and Doomsters, I pointed out that it's easy for small countries to be rich. Here's a pretty striking table making the point:

You will see that the estimates can vary quite a lot from one guessing institution to the next (see Taiwan, for example). But it's pretty clear that many small places display economic performance well in excess of Germany's and even, in quite a few cases, the USA. 

(Just so that you can calibrate these figures with my earlier post, note that, just off the bottom of this list, we find France, the UK, Saudi Arabia and Korea at 25, 26, 27 and 28, Italy at 31, Japan at 36 and Spain at 37.)

So let me return to the question I ducked in my earlier post. Why should this be?

I think we can discount superior governance skills. Singapore is famed for its meritocratic governing class, but Luxembourg - Jean-Claude Juncker notwithstanding - is not, and Switzerland is famed for being so decentralised as to have no real governing class at all. I pointed out earlier that Belgium, Austria and Iceland are not widely regarded as models to emulate, and we in this country, for linguistic and historical reasons, are fairly well aware of flaws of the Irish and Australian ruling classes. I'm prepared to be persuaded that Liechtenstein has a peculiar genetic advantage in statesmanship provided by its isolated mountain location, but I'm not holding my breath.

Let's start with some obvious points. If you're a small country and have natural resources then those natural resources will be shared among fewer people than in the case of a big country. There's obviously the job of managing the resources and doing the sharing out, but that's not as difficult as some jobs in government. Even Singapore benefits from this: it has many things going for it but one of them is the natural resource that is its geographical location. Natural resources can also include some relatively 'unnatural' ones: Panama is the richest country in Central America (per capita), and the fact that it has a big, man-made canal running through it is not irrelevant to that. 

Moreover, some features of a country behave like natural resources. There are a few tax haven-y kinds of country in the list (Ireland included, it's fair to say). Having a combination of secure property rights, internationally legible laws and low taxes can work pretty well: it can ensure a steady stream of company registrations and the odd bit of corporate administration, tax work, litigation and so on, all of which is small beer for the companies or mega-rich individuals involved but adds up to a decent living for the small number of citizens who get to take their cut. 

As I say, that's all obvious. Let's think of a few, more interesting, reasons.

I would guess that small countries are free riders on big ones in certain respects. Take security infrastructure, for example. Belgium and Luxembourg simply don't need to defend themselves, and that frees up money for other things. Germany has also been a bit of a free-rider on defence spending, while France and the UK have come closer to pulling their weight. One might think that the likes of Germany, Belgium and Luxembourg have raked in the GDP points that come from making more butter and fewer guns (or nuclear weapons). One might regard this sort of thing as a bit like insurance: if you save your money on insurance then you will be richer, unless and until some disaster strikes, but if you're lucky in avoiding disaster then you can save a lot of money.

Small countries can get a bit of a free ride on other 'overheads'. France has a certain position in the world as the leading Francophone country, and that comes with obligations to maintain the Académie Française, the benefits of which are felt by French-speakers across the world. You think that's a silly example? Maybe. But it's illustrative of a wider point. A 'big' country is, one way or another, obliged to maintain a diplomatic infrastructure, 'world-class' or 'leading' universities, first-rank cultural institutions and the back-up infrastructure to support them, the odd World Cup or Olympic Games, well-funded medical and research institutions, and so on - all of which are things that have spill-over effects for the rest of the world. The best place for the Olympics to be held is France: we can get there easily and French taxpayers pay. But sometimes we have to grin and bear it ourselves.

These overheads mount up. Think of the covid vaccines. The Oxford/Astra Zeneca one was in large part the result of the UK being the sort of place that researches vaccines - a big, serious country. China and Russia - both per capita poorer than the UK - also had their own vaccines, again because they are big, serious countries. The benefit of these vaccines flows to the whole world (quite rightly, no doubt) but is a net cost to the big country from which they originate: the Astra Zeneca vaccine was not sold to make a profit, for example. Meanwhile, no one expects Luxembourg to join in playing the vaccine-development game. I'm sure Luxembourg pays its share of global or international health projects, but its own population is free to devote themselves to making money running company registration services or what have you. A really big country, like the US, can carry this overhead more easily than a moderately big country like the UK or France.

Small countries can outsource a lot of heavy lifting in this way. Jersey and the Isle of Man can rely on the UK for things like specialist medical care and higher education. They pay, one assumes, but surely only the marginal costs, not their share of the total costs. And does the NHS really charge them full whack? 

Finally, two more speculative thoughts.

First, political overhead. I live in London, which means that my directly- and indirectly-elected representatives include: a couple of local councillors, the mayor of my borough, a geographically-elected London Assembly member (plus a share of the rest based on party lists), the chair of the London Assembly, the Mayor of London, my MP and the Prime Minister. That's really quite a few people. Do we need all of them? And all their hangers-on and administrative staff and so on? 

I wouldn't be surprised if, based on present technologies and social conditions, there are discontinuities of scale in countries: perhaps there's a good efficiency in the very very small countries, another zone of efficiency at about 5m people (Denmark, Finland, Norway, Ireland, Singapore), and another one at about 10m (Austria, Sweden, Belgium; also: London, Paris), but the overhead you need for running the 50-60m-level countries (UK, France, Korea) is quite burdensome, and could just as well be used to run countries at the 80m size (Germany) or even countries bigger than that.  

Second, grands projets. A physically small country makes big, high profile, front-page-of-the-international-newspapers infrastructure projects impossible. That's probably a good thing. The amount of national pride that goes into these things makes it highly likely that they are a waste of money. 

So, for example, it's quite plausible that high speed rail is a bad use of a country's money. Japan and France are famous for their very good fast trains - and neither is a posterchild of economic vitality. Even Italy has a 400 km/h train. Is that a good use of their money? Germany is a great big country, bigger than France or Italy, but its trains run at something more like the speeds we are used to in this country. Maybe the Germans are on to something?

The US, by contrast, is famous for poor public infrastructure, including rubbishy public transport - and yet equally famous for being economically vibrant. Take California, for example: has the population of Poland and an area between that of France and the UK. If California were a European country then it would be bound to have a few high-speed rail lines across it; in reality, being in America, it has one under construction. The US has just one high-speed rail service - Acela - and that averages just 66mph between New York and Boston.

It's easier to think of French vanity projects of this kind than British ones: minitel, TGVs, nuclear power (ok, that was a good one) and the big glass pyramid outside the Louvre. But I suppose we've got "smart motorways", HS2 and the national sport that is arguing about whether to have a third runway at Heathrow, so we try our best. 

I suppose the right attitude to adopt to all this is to make the most of the glorious wastes of money that our country finds itself producing. San Marino and the Cayman Islands are terribly rich and all very nice, I'm sure, but do they have that great big section of Waterloo Station that was used for the Eurostar for a while? Or a London Assembly member elected on the D'Hondt system? Thought not.   

Hanania is right - and his point holds more widely

This post from Richard Hanania has been getting a fair bit of coverage. What he says, at much greater length than I need, is this. While right-wingers in America might not like the US media, and while it's a bit crazy on a few hot topics (you can guess which ones), it's actually pretty good. Not only do the likes of The New York Times turn out large volumes of broadly accurate and interesting stuff, but also they have a commitment to fairness and impartiality which, if not always met (who is perfect?), is at least a lodestar they navigate by in good faith. That a good thing in itself and it also produces the result that the media's output is much better than any nakedly partisan alternative is or would be. 

All true. But there's a bigger point. (More below the break.)

Come, muse, and sing of New Malden!

There's a funny Fry and Laurie sketch in which an ordinary suburban father reveals to his ordinary suburban son that the son is, in fact, the Chosen One who must defeat the evil Pewnack the Destroyer with the aid of the multi-bladed knife Berwhale the Avenger. The son, who had always secretly suspected something of the sort, is then persuaded to move to yonder town of Saffron Walden to wait until the fourth moon of Trollack rises above the Cylinder of Eyelass, and in the meantime to take a job in a canning factory. The sketch ends with the father reporting to the mother that their son seems to have swallowed the story - they thought he would never move out.

There is an online contingent who reminds me of the son. Although they are, in reality, the children of ordinary suburban parents, they feel a great and apocalyptic moral crusade upon them: they are chomping at the bit to pick up Berwhale the Avenger and destroy the evil Pewnack (otherwise known as the Selfish Generation of Boomers) which, they feel, is all that stands between them and a promised land of affordable family homes in south east England. 

What we need to do, perhaps using a trick similar to that played by Stephen Fry, is persuade them to move, not to Saffron Walden, but to its even more prosaic rhyme-mate, New Malden.

Let me explain. An MP made an ill-judged attempt at populism the other day, contrasting "your £500,000 house with a drive in Surrey" with Northern terraces. The internet duly resounded to plaintive cries of "where are these £500,000 Surrey houses with driveways?" Well, I thought I'd have a look for them. It turns out that there are in fact quite a few houses with driveways in Surrey which can had for half a mill. Click that link and you'll see that pretty much all of these houses have driveways. Now I'm sure that the picture in everyone's mind prompted by the MP's tweet is more along the lines of an old Georgian rectory in the gin and jaguar belt of Surrey with an in and out driveway rather than a modest 30s semi with off road parking, but the point still stands: these are decent family homes at which you can charge an electric car.

But now let me tell you a bit about New Malden. It's a London suburb: Surrey in address but London zone 4 in public transport terms. It has good trains to Waterloo (about 8 trains an hour at peak times, almost all of them with a 26 minute journey time). There are lots of primary schools and even some grammar schools in the borough (which bears the distinguished title of the Royal Borough of Kingston upon Thames). It's internationally famous for its Korean community, so it's got some diversity and good restaurants too. Wikipedia says that Stormzy lives there, although I feel that may be stretching a point. 

Now, leaving Stormzy aside, I haven't got any great claims to fame for New Malden. If he were real, I doubt Pewnack the Destroyer would be seen dead there. On the other hand, I'm pretty sure that Reggie Perrin was eleven minutes late because his train was stuck there once. But it looks like a good place to settle down and bring up children. Parks, playgrounds, a High Street and so on, and all within easy commuting distance of London. That, as I understand the online Chosen Ones, is the promised land.

And here's the thing: you can buy a three bedroom house with a driveway there for half a million pounds. Honestly: the dream of a house with a garage and a garden in London with a decent commute is alive and well, and it lives in New Malden. Here's one that sold earlier this year for £460,000. It's not the world's most beautiful house. But if you wanted 1200 square feet of indoor space and a 55ft garden in zone 4 then it could have been your house, for £460,000. And there are a few more of similar ilk on the market. Again, these are not houses to make you swoon - they are not even the fashionable Victorian terraces I talked about before - but they are homes which will accommodate a young and growing small family. They are probably not unlike the drab homes that the Boomers bought once upon a time and which now make the Chosen Ones green with envy.

So pack up your multi-bladed knives and lie low in New Malden, biding your time until the day of the Great Reckoning. In the meantime, you might end up making a quiet suburb a little more lively - and a hefty profit on your deposit too.

High status beliefs: is Brexit the Britten of politics?

The other day, I tentatively outlined a quantity theory of high status beliefs. Having re-read that sentence, I appreciate that it is rather niche. I'll put the rest below the break.

On Wren-Lewis on cuts

Simon Wren-Lewis has written an interesting blog post essentially in support of the proposition that, as he puts it, “It makes no sense to keep shrinking the state when you don't change what the state does.”  I disagree and hope to show why.

Are the UK's cities poor because they are new?

I apologise for presenting a rather inchoate theory, but I am sure I am onto something. Here goes.

This graph, produced by the ever-interesting Tom Forth, has been doing the rounds. 

I don't know if you can see it properly but I hope you get the idea. London (the upper-most orange bar) is rich, but all of the UK's other cities are relatively poor by European standards: even the likes of Warsaw, Prague, Budapest and Bilbao can be found in the mass of blue between London and Manchester (the second orange line). 

You could quibble about the numbers (does Poland do surprising well out of the PPP calculation here?) but I would suggest you don't bother: if you've been there you will know that Paris and London are richer than Valencia or Liverpool. (NB. Forth has taken Dublin out because the numbers are misleading and he later removed Oslo for being non-EU. Really don't worry too much about the details.)

Why should this be? Well, perhaps it's the lack of light rail systems or other 'investment' etc etc etc. I suspect that is the direction Forth would like to go in. But I'm inclined to think there's something more fundamental. 

Here's the list of cities from top to bottom in what I hope is a more legible format:

The point that immediately struck me is that most of the European cities are ones that were successful before the Industrial Revolution, while most of the British ones were not. 

Perhaps you're already persuaded of the point - but perhaps not and need proof. So let's start with France, our sweet enemy and the eternal comparator of England/the UK. The French cities in the list are: Paris, Lyon, Toulouse, Nice, Marseille, Bordeaux and, between Leeds and Birmingham, Lille. Here's Wikipedia on Early Modern France: "Paris was one of the most populated cities in Europe (estimated at 400,000 inhabitants in 1550; 650,000 at the end of the 18th century). Other major French cities include Lyon, Rouen, Bordeaux, Toulouse, and Marseille." What about Lille and Nice? Well, they weren't French at the time! But they were well-established towns.

Now let's look at the biggest towns in England in the 17th century (so far as we can tell from the hearth tax of 1662):

As Wikipedia points out, "Most notable from a modern viewpoint is the fact that Manchester, Birmingham, Leeds, Liverpool and Sheffield do not make the top thirty, whereas within around 100 years they would become England's largest provincial cities". Indeed, Newcastle and London are the only cities on both the chart at the top of this page and this list of 12 cities, while the overlap in France is plainly much higher.

Here's some corroboration of my theory that British towns are 'newer' than French ones. It comes from a proper academic paper: 

I got that from here, courtesy of Tyler Cowen, who drily comments, "These days, the French model is looking somewhat better, as Toulouse has held its ground more readily than has Liverpool." I'll come back to that comment below, but for the moment let's just note that there is a striking contrast here.

So that's France. Now, let's take the cities in the list from Low Countries: Amsterdam, Brussels, Antwerp, Rotterdam and The Hague. I'm sure you don't need to be told that Amsterdam and Brussels were well-established before the Industrial Revolution, but you might not know that Antwerp was of comparable size to them in the Early Modern period. The Hague became the permanent seat of the States of Holland in 1588, so it was a big deal then. (Note also that these are OECD standard definition 'cities', so not quite what you might expect - "London" includes Sevenoaks, for example, and "Leeds" includes Bradford: "The Hague" includes Delft, to give you some idea of the antiquity of the urban area we are talking about.) I grant you that Rotterdam is a late arrival, interestingly mostly after 1872 (it's more of a post-industrial revolution city), but I would suggest that we see a similar pattern to France: pretty much all the successful cities were successful before the Industrial Revolution.

Now I freely admit that I haven't gone down the whole list, but I don't think I need to: if you simply cast your eye over it you'll see cities as old as Athens and Rome. However, in my ignorance I had a quick look at a few that concerned me.

 

- Milan? Maybe not as old as the likes of Florence or Naples? Ah but did you know that the "The Great Plague of Milan in 1629–31 ... claimed the lives of an estimated 60,000 people out of a population of 130,000"? A population of 130,000 in the 1600s! That's twice the size of Norwich at the time. 

- Turin? "From 1563, it was the capital of the Duchy of Savoy, then of the Kingdom of Sardinia ruled by the House of Savoy, and the first capital of the Kingdom of Italy from 1861 to 1865". That's from Wikipedia, which tells me that it had about 90,000 inhabitants in pre-modern times, i.e. 1.5 Norwiches.

- Gothenburg? That's a city which, it turns out, "was founded in 1621 and became the big town in the west. The underlying strategy of the Swedish government was to create a seaport in the western part of the country that could function as an important link between Swedish trade and a global market ... 17th century Gothenburg was an internationally oriented seaport with a lot of foreign people living in the town". (In fact, both Gothenburg and Rotterdam made money from pre-industrial far eastern trade.) Note that Birmingham wasn't even one of the 30 biggest towns in England while Gothenburg was already a cosmopolitan entrepot.

- Stuttgart? Sure, it's a big car-making place now, but it was a big deal before the Thirty Years War, with Dukes of Württemberg, Hapsburgs and so on involved in complicated ways.

Katowice is an industrial revolution town, as is Mannheim-Ludwigshafen (BASF's headquarters), so it's not just the UK that has large new cities. But the UK is surely unusual in the extent to which its largest cities are the product of the Industrial Revolution.

So what do we make of all that? The Brit-booster approach - and, frankly, not a silly one - is to say that the UK has done the amazing job of creating, essentially overnight in European historical terms, metropolises with wealth to rival ancient and historic cities such as Lisbon, Dresden, Bordeaux, Athens and Seville - while sparing its own equivalent cities, such as Oxford, Cambridge, York, Norwich - from the depradations of industrialisation. Good job UK!

There's something to that. Many places that left industrialisation to the new boys are going strong. Here are the 20 least affordable (which I am taking as a proxy for most sought-after) towns in the UK according to the Halifax in 2021.

1. Winchester
2. Oxford
=3. Truro
=3. Bath
5. Chichester
6. Cambridge
7. Brighton and Hove
8. London
=9. St Albans
=9. Chelmsford
11. Salisbury
12. Exeter
13. Leicester
14. Norwich
15. Bristol
=16. Southampton
=16. Canterbury
=16. Gloucester
19. Worcester
20. Cardiff

We see a number of towns famous for Georgian (Brighton, Bath) or older architecture, and several Roman towns on the list (all those "chesters", plus St Albans). There's also a fair degree of overlap with the list of biggest towns in 1662: Oxford, Cambridge, Canterbury, Norwich, Bristol, Salisbury, Exeter, Worcester. By contrast, Bradford, Newcastle, Glasgow and Liverpool all feature in the equivalent "most affordable" list.

To be absolutely blunt, we all know that nice old towns that escaped the Industrial Revolution are lovely places to visit or live. Durham and Stamford are small, pretty and expensive: this house is very lovely, but it's a 6 bedroom terraced house in Stamford, not a mansion in London, and yet it went for £2.8m (actual sale price). This 5 bedroom house in Durham has an asking price of £2.75m. These are big sums (and a bit different from nearby Newcastle or Peterborough).

So far, I've been talking about England in 1662 and comparing it with the UK today. That's not right: what about Scotland? Well, the picture is similar. The Scottish city in the list is Glasgow, another Industrial Revolution boom town, while Edinburgh, the charming combination of old Old Town and Georgian New Town and the biggest town in Scotland in the 1755 census, doesn't make the cut. A quick check on Rightmove says that Edinburgh is at least 50% more expensive than Glasgow. I accept that Glasgow was a well-established town before the Industrial Revolution, like Newcastle in that respect, but I don't think I'm stretching a point in saying that its post-Industrial Revolution history predominates.

Let's assume I'm right. On the Continent, for whatever reason, they have a number of big cities which are the equivalent of what would have happened if York or Norwich had taken Manchester's place in our industrial history, while our cities are Milton Keyneses or Dubais, shiny newbuilds or johnny-come-latelys. So what?

Let's go back to Forth's concern and Cowen's comment. The UK is richer than, say, Poland. So surely it's a bit odd that the UK's big cities are no richer than Poland's? Why should this be? Is there any reason to think that the fact that the UK's cities are "new", while the rest of Europe's are "old" is relevant?

This is a little speculative but I can think of two reasons why it might be relevant.

(1) The UK's cities, except London and Newcastle, are boom towns created for a boom that busted. They are like mining villages after the mine has closed. Why are they there at all? We know why towns that were big before the Industrial Revolution exist: something to do with ports, rivers, (Roman) roads, prevailing winds and all the rest of it. They had a raison d'être before industrialisation and it's not surprising that they still have one. But the UK's new cities? Aren't they just European Detroits? Big congolomerations of built stuff in the wrong place? The big German industrial cities look terribly rich and successful now, but assuming Germany lives to see a post-industrial future, won't the likes of Mannheim-Ludwigshafen have to endure Detroitification as well?

(2) Or try this idea instead. Let's look at where people want to live in a post-industrial age. The capital city of any country always has a pull and it's not surprising that we see many capitals in the list above (e.g. Berlin, which was transformed by the Industrial Revolution). But apart from that people will, surely, tend to choose nice places to live - and a city with a pre-Industrial Revolution history is, other things being equal, just, well, nicer. I am a big fan of Victorian architecture! I've got nothing against the likes of Leeds. But Oxford has some pretty good Victorian architecture - plus some rather older examples too. Perhaps that's an unfair comparison, but I'm surely getting at something. A European city - here or abroad - that consists of a mediaeval or early modern core, surrounded by 19th or 20th century developments, has the edge on one that only has a 19th century core. People go on holiday to Vienna, Amsterdam, Munich, Prague or Barcelona - they will get on a plane and spend hundreds of pounds just to visit one such city. The second cities of the UK simply don't have that appeal. There's a cheese called "Cathedral City" but there's no cheese called "Industrial Conglomeration".

All in all, I strongly suspect that pouring investment into the Industrial Revolution towns of the UK is not going to make the same returns as similar investments in, say, Amsterdam or Munich. Wealthy, successful middle-class people can make money in Amsterdam and Munich, and they want to live there. Their equivalents in the UK either struggle to make big money in cities built around industries that don't exist, or perhaps, if they can make big money, they prefer to live in Stamford or Durham, the Cotswolds or Rye, Royal Leamington Spa or Royal Tunbridge Wells, or ... 

Perhaps that means that there is a fantastic opportunity for the UK? Surely we could just take the likes of Norwich, York, Oxford, Cambridge, Ludlow, Chester, Canterbury etc and turn these fantastically appealing towns into the cores of much larger but nonetheless still appealing cities with Turin/Antwerp/Prague levels of wealth and income? All we need to do is build them out in the same way that Turin etc have been built out. 

Possibly. But that wouldn't be easy. Take York, for example, a town so rich and lovely that, according to The Economist, it seeks to avoid any new developments at all. And who could blame it? It's been a long time since anyone could be confident that major developments on the edge of a cathedral city would result in a more attractive or pleasant town at the end of the process. 

But what's the alternative? As ever with these things, it's a bit of a mix and match. The UK has moved its centres of (non-London) activity around before, from the Suffolk wool towns to northern industrial cities, for example, and no doubt it can do so again. There are geographical constraints on some of the smaller cities that people want to live in (Bristol and Brighton spring to mind), but towns that already have a dense Georgian street pattern can probably grow in numbers, at least to some extent, without growing in physical footprint. Other towns may find developers more sympathetic to their needs than those York has seen recently, and they will be able to grow outwards: Cambridge might be an example. Some big Industrial Revolution cities will find new life in the post-industrial age: Manchester may become a media-and-university powerhouse, for example, and it seems to do well in IT services too. 

But perhaps we also need to appreciate that some cities are just in the wrong place. There's a danger that 'investing' in some towns simply amounts to bribing their inhabitants to stay away from the places where they could be more usefully or happily employed. Are we subsidising the maintenance of industrial revolution heritage theme parks at the expense of a post-industrial future based in places that people really want to live?

And there are positives to people moving on too. In the 14th century, King's Lynn was England's most important port. It's not now. That has left King's Lynn with some lovely things - Hanseatic architecture, the largest chapel of ease in the country - that make it a nicer place than a purely modern town of an equivalent size. Maybe the glories of Bradford Town Hall will one day grace a lovely little village that people who live and work in the thriving hubbub of Hebden Bridge Spaceport visit for a quiet weekend? Is that not at least as enticing a vision as the idea of pouring taxpayers' money into light rail schemes?