MM2

Code for exploring Benjamin Plener Cover's Two-Party Structural Countermandering (aka "MM2").

Base Apportionment

The HH_Apportioner class implements the Huntington-Hill method used today to apportion representatives to states for the House based on a decennial census. It takes a census data as input: a list of dicts read from a CSV by DictReader (using read_csv()). Examples are in the data/census/ directory. To apportion representatives to states based on a census, call the assign_first_N() method with a 435.

app = HH_Apportioner(census)
app.assign_first_N(435)

The class generates a queue of priority values sorted in descending order, initially 100 entries per state. You can continue to allocate seats to states, by calling the assign_next() method which returns the # of the House seat, the priority value, the state that gets the seat, and the seat # for the state.

Augmented Apportionment

The MM2Apportioner class wraps the HH_Apportioner class and adds functionality for apportioning addition (party) list seats to states, again based on a decennial census, and then allocating the list seats to parties based on the results of an election. This can be used in two modes. To apportion nominal & list seats to states based on a census:

app = MM2Apportioner(
    census, None, list_min=list_min, total_seats=size, verbose=verbose
)
app.apportion_seats()

The apportion_seats.py does this and writes results into a file named "{census year}_census_reps_by_state({size},{list minimum}).csv" into the results/ directory, where:

• {census year} is the year of the census data
• {size} is the total number of seats to apportion (both nominal & list), and
• {list minimum} is the minimum number of list seats guaranteed for each state (0 or 1)

To allocate the list seats apportioned to states to the parties based on election data:

appr = MM2Apportioner(
    census, elections, list_min=list_min, total_seats=size, verbose=verbose
)
app.apportion_and_assign_seats()

The assign_seats.py scripts does and writes reports into the results/ directory:

1. {election year}_report({size},{list minimum}).txt summarizes the run.
2. {election year}_reps_by_state({size},{list minimum}).csv summarizes nominal & list seats by state. [//]: # (3. {election year}_reps_by_priority({strategy}).csv shows the seat-by-seat assignments.)

where:

• {election year} is the year of the election data, and
• the other variables are as above

The logic of these two processes is described in Logic.

Our explorations of alternative MM2 apportionment strategies are described in Explorations.