On Tuesday, President Trump signed a memorandum — sure to be challenged in court as violating the Constitution — intended to bar undocumented immigrants from being counted for congressional apportionment next year.

This move is audacious, given its dubious constitutionality. But it is also unsurprising, given the Trump administration’s failed attempts to introduce a citizenship question to the U.S. census for the first time in decades and the inherently political nature of the census because of its role as the unique instrument of proportional representation of the states in the House of Representatives.

The question of who should count is paramount to the issue of how representative power should be distributed. The first census counted free people and enslaved people, albeit only as three-fifths of a person, but excluded untaxed Indians. The question of whether and how to count foreign citizens living in the United States dates back to the mid-19th century. Two questions underlie these debates — which areas of the country, urban or rural, Northern or Southern, would have political power and what part of the population deserved to be represented, because race and space were inseparable. The Trump administration’s effort not to count undocumented immigrants is nothing less than an effort to redistribute political power, one that calls to mind a particularly fierce battle over the 1920 census that highlights the role of these broader fights.

Like in 2020, the 1920 Census was, from the start, highly charged. In January 1920, which was then the census month, the Department of Justice’s decision to raid thousands of homes of presumed radicals in the days when the census takers were starting to canvass the population affected the accuracy of the count. The famous “Palmer Raids” not only resulted in thousands of arrests and hundreds of deportations. They also created fear and mistrust of government officials among immigrants that resulted in incomplete counts of immigrant populations across the country. A knock at the door could be harmless census takers or federal agents seeking to arrest and deport “radicals.”

Historians think that this led to an undercount of the nonnative-born population, slightly reducing the population of the states where they lived in the greatest clusters and more importantly identifying immigrants in general as a menace to the body politic. Yet despite these likely undercounts, the census accurately showed that for the first time in its history, the United States had a majority urban population. The rapid growth of the industrial states in the North due in large part to recent international migration — especially of the “new immigrants” from Southern and Eastern Europe — was poised to have political consequences. In redrawing district lines on the basis of the census count, the Northern industrial states were about to take congressional seats from the more rural Southern states.

In December 1920 and January 1921, after the results of that year’s census were in, therefore, a bitter dispute took place in the House of Representatives over apportionment and how the census could be used to deliver fair representation. Southern states and rural districts were not only threatened numerically by a census that showed an increasingly urbanized population. The census might also contain proof that black voters were being suppressed, something that could diminish congressional power of states engaging in voter suppression.