New article from Ruoxi Li in PS. Abstract:

In the latest round of the apportionment of the US House of Representatives following the 2020 Census, the State of New York lost a seat by an extremely small margin: if a mere 89 people were added to the state’s population of 20 million, the state would have kept the seat. Political observers pointed to the tendency of the US Census to undercount minority and immigrant populations as the primary culprit. However, New York’s seat loss is as much an issue of apportionment as it is of counting. The current apportionment method used by the federal government, Huntington-Hill’s method, is biased against more populous states such as New York. If an alternative apportionment method were used, such as Webster’s method, New York would have kept the seat. This article discusses four historical apportionment methods: Hamilton’s method, Huntington-Hill’s method, Jefferson’s method, and Webster’s method. These methods are evaluated against the three criteria of within-quota, consistency, and unbiasedness. The article shows that Huntington-Hill’s method has produced biased apportionment results in eight of nine apportionments since its official adoption in 1941. The article concludes with a recommendation for replacing the current apportionment method with the only unbiased divisor method: Webster’s method.