This post is to explain why comparative policy analysis must start with sovereign states, rather than their subnational entities. This includes urbanism, despite the anti-national orientation of some urbanists. In particular, all of the following claims that I routinely have to deal with are incorrect:

• The claim by Americans that the US must actually be compared with the EU, and its states with EU member states, on matters of economic and social performance.
• The more fundamental claim by Americans that American states, subcultures, ethnic groups, and other subdivisions are so different from one another that they have to be compared with European nation-states and the US writ large is too diverse to be so compared.
• The claim by urbanists that metropolitan areas are a more natural unit of analysis than states, for example the transnational Blue Banana as opposed to the EU member states (and UK) that it overlaps.

Policy and politics

Politics in virtually all first-world democracies is nationalized. Political parties are nationally unified, and the president (if executive) or prime minister is a well-known national figure. Regional politicians can have special features relevant to their region, for example big city mayors tend to be more cosmopolitan than is the norm for their parties but are still understood to represent the party and its governance. Even in countries with a split between national and provincial parties, like Canada, there is enough overlap that one can talk about the Doug Ford faction of the Conservatives and about the factions Pierre Poilievre is appealing to even though Ford runs Ontario and isn’t a federal politician.

None of this exists across national boundaries. The EU has EU-wide party umbrellas so that one can identify green parties, social democratic parties, right-liberal parties, Christian democratic parties, various flavors of extreme right parties, and so on, but it’s almost unheard of to attack a party in national politics based on what another party in the same umbrella group in a different member state. Even on fundamental issues, parties that coordinate closely may differ: there’s an umbrella group of agrarian Nordic parties, whose Norwegian member is strongly against joining the EU and whose Swedish and Finnish members are not particularly euroskeptical within the Union (its Danish member is the main center-right party and supports joining the euro).

Different enacted policies can be used as a mirror with which one judges one’s own country’s politics. Thus, in Sweden, people extensively look at what other Nordic countries do and try to learn lessons, or, more often, pick and choose examples to justify their preconceptions of what Sweden should do. Finnish education was very popular in Sweden in the 2010s as an aspirational comparison. In Finland, the Green Party, Vihr, looked at the shutdown of nuclear power in Germany without adequate renewable replacement, and decided against backing a similar policy for Finland; the party supports maintaining the nuclear plants for now. But it would be unthinkable in Finnish politics to attack Vihr for a policy backed by the German Greens, or even by the Swedish Greens, Mp; for example, Mp voted against joining NATO, but this has not led to any attacks on Vihr, which voted in favor and which is in support of rearmament to deter Russia.

Within countries, this is completely different. CDU cuts ads attacking the governance of Berlin under all-left coalitions for national consumption. Most people on the German center-left speak with disdain of the Lower Saxony gang, the SPD elite leadership drawn from the state with ties to heavy industry, general political inertia, and a pro-Russia foreign policy (Schröder was mayor of Hanover before he became chancellor, and other 2000s-10s SPD elites from the state were warm to Russia). In the United States, moderate Democrats across the country find themselves having to denounce Zohran Mamdani, who is not in national politics and doesn’t so far have much of a path to go into national politics.

What the state does

The OECD Fiscal Decentralisation Database shows the proportions of taxation and spending at each level of government. Some states are atypically centralized, for example the United Kingdom. Others are atypically decentralized, for example Australia or the United States. But even in the United States, taxation and spending are majority-federal, and the biggest state spending program is typically Medicaid, heavily federally regulated in law and politics.

Meanwhile, at the EU level, the share of the budget is so small it might as well not exist. The social insurance, health care, and military spending that dominate spending are entirely at the member state level, or even devolved further. (Nordic health care is managed at a regional level and not at the state level; Finland has 21 well-being regions for health and welfare.) The EU has a very large role to play in the sort of regulations nobody notices and are not subject to mass politics, for example coordinating health insurance and education reciprocity between member states, or coordinating rail rolling stock regulations so that the national homologation processes are mostly the same and the same trains that run in Germany can run in France and vice versa.

Not for nothing, the infrastructure system shows some hard borders. Within EU member states, intercity rail can be fast and convenient. Most major provincial cities in France are connected to Paris by TGVs averaging more than 200 km/h end-to-end. German trains are slower, but the national network has tight integration as far as can be pursued on such a complex system, and further investment aims to tighten this integration, to enable somewhat higher average speeds. None of that exists across borders. The only cross-border all-high-speed system in the EU, from Paris north to Brussels and thence Holland, is beset by high fares, designed for business travelers only, nothing like the more popular model used by SNCF domestically; the most recent data I have seen has international TGVs and Eurostar averaging twice the fares per passenger-km of domestic TGVs. States build high-speed railways and run trains on them; the EU passes passenger rights regulations to deal with some cross-border frictions.

In the United States, it’s hard to exactly compare the situation, since the only trains that we would consider bad but barely passable are the Northeast Corridor and a few branches; the other trains might as well not exist. But the trains that do exist do not meaningfully respect state boundaries, because American geography developed as a single polity and not as 13 or 50. Even state-subsidized corridor trains on Amtrak are often cross-border, such as the Michigan trains to Chicago, or the Illinois-subsidized trains to St. Louis. The highway network, meanwhile, operates on a single national standard, and so does the freight rail network, without the labor frictions seen at some intra-EU borders that frustrate the use of rail freight at long distances.

Economic and social outcomes

Because politics is national and most services are provided directly or indirectly at the national level, the economy is national as well. This can also be seen in economic growth. Megaregions are invisible in it, and national growth matters more than anything. For sure, there can be trends within a country, with relative decline or growth in various regions, but these don’t really align with any story of globalized megaregions. In Germany, one can see the convergence of East Germany to West German living standards, but otherwise, megaregions like the Blue Banana are irrelevant – Hamburg, firmly West German and yet outside the Blue Banana, doesn’t behave differently from German metropolitan areas in the Blue Banana like Frankfurt. And in Italy, the economic stagnation from the 1990s to 2020 affected both the North and the South – there was no further divergence as the Blue Banana model might have predicted.

The same can be said of social outcomes. I’ve repeatedly made Americans on social media angry by pointing out that the United States has low life expectancy. They keep retorting that actually the problem is drag from the Deep South, and New York and California are doing fine, with life expectancy about two years more than the national average. And yet, life expectancy about two years more than the national average is rather common in rich regions with selective migration, even in European countries with even access to health care, like France and Spain.

Culture and migration

People in a particular country or region are immersed in its media and learn in detail about its differences. When researching the Swedish case, I kept finding these intra-Nordic comparisons that focused on differences in road funding details between Sweden and Norway, differences that from outside Scandinavia are invisible (in fact, the construction cost history of Sweden, Norway, and Finland is largely the same, with Denmark being only somewhat different).

This can lead to a kind of insularity, in which people are too aware of their own country’s internal diversity and insufficiently aware of others’. Thus Americans can recite details of how different regions or ethnic groups in the country behave, and even pronounce certain cities European, cities that lack pretty basic aspects of common European culture (say, association football). This is where the mentality among Americans that the interregional cultural differences in the US are more comparable to differences between rather than within European countries comes from. They are only vaguely aware of German stereotypes involving Bavaria, a state that views itself as so special that it’s had separate political parties since 1918, or of England-Scotland and Wallonia-Flanders differences that are in some cases greater than interstate differences in the US (the American education system is broadly the same regardless of state; England and Scotland don’t even have the same length of undergraduate education).

The United States, after all, is a country with extensive internal migration. The Census Bureau’s 2024 estimate has 15.7% of people resident in the United States born abroad or in Puerto Rico, and 43% born outside the state they live in. Among adults, a slight majority, 51%, were born outside the state they live in. Immigration numbers in Europe look rather like the immigration numbers to the United States, and not at all like the interstate migration numbers within the United States. EU mobility is rising but still so far from intra-US mobility that the comparison is silly.

This is why, in mass transit, the United States has a national class of planners, consultants, agency managers, and staff. Managers move between major systems easily, and planners in California are expected to be aware of what is happening on the East Coast (in the other direction, it’s less expected, because of core-periphery dynamics). The assumptions made in Seattle, Los Angeles, Boston, and Washington are broadly the same. The assumptions made in New York have a few idiosyncracies coming from New York feeling special, but they’re still within the same constellation of what people think is obvious elsewhere in the United States, and overall the differences aren’t larger than those between German cities or between French ones.

Those assumptions not only are different in Europe, but are different between countries: the British, French, and German ways of building rapid transit are different, and there isn’t even a unified Nordic way but rather different paradigms in different Nordic capitals (and yet the costs and procurement systems are the same). This cascades down to the level of activism: German environmental advocates are sour on U-Bahn expansion, viewing it as part of a cars-and-trains paradigm and preferring trams; in contrast, anti-subway activism in green politics is uncommon outside the German-speaking world.

At most, one can say the United States is comparable in its internal diversity to the combination of Germany and Austria, in some aspects. One cannot compare it to Europe writ large, nor can one compare its states to EU member states rather than to subnational divisions like German Länder. Americans in their solipsism try to convince themselves there is no need to study the rest of the world in detail, but merely to collapse it into macroregions like Europe or Latin America, and to analogize entire nations with their diasporas in the United States. But it’s not meaningfully possible to understand the world at this scale – macroregions are real but analysis of policy, including transportation but also the usual economic and social topics, has to be down at the level of sovereign states first with no illusion that a province in a large country is comparable to a smaller country.