Getting to Denmark.

On 10 July 1920 King Christian X rode a white horse across the old border at Frederikshøj. A nine-year-old girl from the crowd was lifted onto the horse in front of him. Her name was Johanne Braren, and the photograph of the two of them became one of the best-known pictures of the reunion.

The people lining the road had been citizens of Prussia and then of the German Empire for 56 years. Since 1888 their children had been taught in German, yet none of them had moved. A war in 1864 had moved the border north over their farms and towns, and in 1920 a vote moved it back.

What had 56 years of German rule done to them?

That question is a small version of a large one. Political scientists have argued for decades about whether the habits that make some societies work are inherited from the deep past or rebuilt by each generation’s institutions. Denmark is the case everyone reaches for, and here I follow its southern border through two changes of rule, three kinds of evidence and a century and a half, to see how long history lasts.

An interviewer reads from a card: “Generally speaking, would you say that most people can be trusted, or that you can’t be too careful in dealing with people?” Zero means you can’t be too careful; ten means most people can be trusted. In survey after survey the Nordic countries answer at the top of the world’s distribution, and Denmark sits at the top of the Nordic countries.

Trust lowers the cost of almost everything a society does together, because people who expect others to cooperate need fewer safeguards and less monitoring. Studies have linked high trust to faster economic growth, lower corruption and more responsive democracy.

So the rest of the world has taken notes. The Economist called the Nordic countries “the next supermodel” in 2013. In 2016 Bernie Sanders campaigned for the American presidency holding up Denmark as a model. Development economists and political scientists even have a phrase for the ambition of building a state that is prosperous, democratic, peaceful and honest: getting to Denmark, popularised by Francis Fukuyama.

The obvious next question, and the one I start from, is how Denmark got to Denmark.

There are two broad answers to how the gap arose, and they have very different implications.

The first says Danish trust is old. It points to centuries of political stability and relative social harmony, to an industrialisation in the late nineteenth century driven by cooperatives of peasants who banded together to buy and share equipment, and to a striking fact about emigrants: descendants of Scandinavians who left for the United States around 1900 are still among the most trusting Americans, close to Scandinavians themselves. On this account trust is a cultural inheritance, carried in families. Some scholars have argued that it explains Nordic success, including the welfare state itself.

The second says Danish trust is recent. It points to the universal welfare states built after the Second World War, to low inequality and low corruption. People trust each other because their institutions treat them fairly and give them little reason not to. On this account trust is made, and the rise since 1980 is part of the evidence.

For governments trying to learn from Denmark, the second account is the hopeful one, because it implies that what Denmark built, others can build. The trouble is that the surveys start around 1980, when Danes were already far ahead. They cannot tell us when the gap opened. For that we need a place where the same kind of people lived under Danish and German rule at different times, and history provides one.

For centuries the Danish crown ruled more than Denmark. It held the Duchy of Schleswig as a fief from the eleventh century and, from 1773, the whole of Schleswig and Holstein, which reached south to the Elbe and the edge of Hamburg. Denmark was what historians call a composite state: one king, several territories, each governed somewhat differently. The duchies were administered in German, by German-speaking landowners, under a Danish king.

Nationalism tore that arrangement apart. In 1848 German speakers in the duchies rose against Copenhagen, and the First Schleswig War that followed killed 14,000 soldiers on both sides before Russian pressure and British mediation ended it. Sixteen years later Prussia and Austria attacked.

Defeat at Dybbøl became a national trauma, and Denmark answered it by turning inward. The poet H. P. Holst gave the country its slogan in 1872:

What followed in Denmark is remembered as the age of associations: a flurry of popular mass movements, voluntary organisations and free church congregations. Politically the country kept moving towards self-government. Parliament won sovereignty over the king in 1901, and broad enfranchisement in 1915 made Denmark a full democracy.

North Schleswig lived through the same decades inside a different state. The German Empire held national elections from 1871 under a formally secret ballot, though effective safeguards for secrecy came only in 1903. Prussia’s separate state elections used a three-class system in which a man’s tax payments determined the weight of his vote. The parliament faced a very strong executive under the chancellor, Otto von Bismarck. Danish newspapers and associations campaigning for reunion were pursued by the authorities. Shortly after 1866 North Schleswigers signed a petition for a referendum on their future; 27,473 people put their names to it. From 1888 the state schools taught in German only, except in religion lessons.

The Danish speakers organised in response. They built “free” congregations, “free” schools and about fifty meeting houses to keep the language alive. That detail matters later: being a repressed minority can also make a community pull together.

The difference between the two states, in other words, was mostly about liberty: how freely people could print, meet, organise and vote. It was not about the quality of administration. Danish officials had been trained along German lines since the seventeenth century, and the Prussian bureaucracy was comparably professional. Nor was it about land: the great estates of Holstein and much of Schleswig were as unequal as those of Prussia. So this is a test of liberal, representative rule, and of what being a minority does to a community. It cannot test bureaucracy or land reform.

Germany lost the First World War, and the peace settlement offered Schleswig a choice. The plebiscite was held in two zones. The northern zone voted first, on 10 February 1920, and the central zone, which included Flensburg, on 14 March.

The beauty of a border that moved twice is that the two tests predict opposite signs. A legacy of liberal Danish rule should push the duchies up and North Schleswig down. A legacy of minority life should push both up, because both spent decades as a minority defending their language against a state run by the other side. So the evidence can tell the two accounts apart.

Nobody asked people in 1871 whether they trusted their neighbours. But they did vote, and voting is a reasonable place to look. Turnout is behaviour rather than an answer to an interviewer, it covers the whole electorate, and in today’s surveys people who trust others are more likely to vote; in the European Social Survey the correlations for Denmark and Germany are about 0.2.

Germany began holding elections to the national parliament in 1871. I take every constituency in the former duchies and every constituency in the neighbouring German states, and follow the difference in turnout from 1871 to 2009, leaving out the Danish-speaking constituencies so that this is German speakers compared with German speakers. In Denmark I do the same for North Schleswig and its Danish neighbours from the first election after reunion to 2011.

The first thing that stands out is that the signs are opposite, just as a legacy of liberal rule predicts. Former Danish rule goes with more voting in Germany, former German rule with less voting in Denmark. If being a repressed minority had mobilised communities, both regions should have voted more. They did not, which also rules out the idea that any change of state simply unsettles people for a while: the duchies changed state too, and voted more than their neighbours.

The second is that the effects fade, and within a lifetime the voting habits that the old state left behind had all but disappeared.

There is a second test, from associations rather than ballots: how many people belonged to a gymnastics club in 1865, county by county across Prussia.

Turnout is only a proxy for the question I started with, which is about trust itself, and for that we need the survey interviewer again. Six rounds of the European Social Survey, from 2002 to 2012, include 850 respondents living in the former Danish duchies and 460 in North Schleswig. That is enough to compare each group with its neighbours on the other side of the old line, in the same country.

For Germany I matched each constituency to the 1864 map and measured how much of it the Danish crown had ruled: 1.1% of Schwerin-Ludwigslust, 50.4% of Lübeck, 100% of all eight Schleswig constituencies. In Denmark the survey records the municipality. The models hold constant age, sex, education, income, religion, minority and migrant background, and the survey round.

An average over everyone can hide a difference among some. If German rule left a mark on North Schleswig, it should be clearest among people born soon after 1920, whose parents had grown up under it, and weaker among their grandchildren. So I allowed the North Schleswig difference to change with each respondent’s year of birth.

In the former duchies the equivalent estimates are small and imprecise, but they point the same way as the turnout evidence: a modest early advantage from Danish rule that wears off. In North Schleswig the oldest generations may still carry a remnant of the German years. It is small next to the gap between Denmark and Germany, and it is gone among people born after the middle of the twentieth century.

Across the three pieces of evidence, both changes of rule moved political behaviour in the direction a liberal-institutions account predicts. Both effects faded within 40 to 70 years. Today, in the regions that changed hands, trust shows no consistent difference from their neighbours, apart from a remnant among the oldest people in North Schleswig.

If these border regions generalise to Denmark and Germany as a whole, then most of the very large gap in trust between the two countries cannot be a direct inheritance from rule before 1920. The direct responsibility lies with what happened after. That is a strong conclusion against accounts that explain Nordic success by a deep and durable culture of trust.

The other Nordic countries point the same way. Norway and Sweden were governed much like Denmark through the nineteenth century, and liberalised gradually as it did, so the conclusion may well extend to them. Finland is the hardest case for the deep-history view. It was ruled by Sweden, then by Russia as an autonomous grand duchy from 1809 to 1917, and in 1918 it fought a civil war that killed about 37,000 people out of a population of 3 million. Its levels of trust today are similar to those of the other Nordic countries.

None of this means history is irrelevant, since the design cannot see indirect effects. High trust and a habit of association may have made it easier for Denmark to adopt proportional representation in 1915, or to take the first steps towards a welfare state in the early 1930s; in Sweden, the early welfare state took over a great deal of work that voluntary organisations were already doing. Institutions built in one generation can carry the habits of the last into the next. What the border evidence argues against is the simplest story, in which Danes trust each other because their great-grandparents did.

This matters now because “getting to Denmark” is still the question that reformers, donors and governments in many countries ask. The pessimistic reading of Danish success is that some countries are simply blessed with a culture that others cannot copy. The evidence from this border points the other way. Within two generations, the people of Schleswig came to vote and to trust like the country they lived in. On this border, trust behaved like a habit that institutions teach, and that a new state can teach again.