Gerrymandering and computational redistricting

Olivia Guest,Frank J. Kanayet,Bradley C. Love

DOI: https://doi.org/10.1007/s42001-019-00053-9

2020-06-28

Abstract:Partisan gerrymandering poses a threat to democracy. Moreover, the complexity of the districting task may exceed human capacities. One potential solution is using computational models to automate the districting process by optimizing objective and open criteria, such as how spatially compact districts are. We formulated one such model that minimized pairwise distance between voters within a district. Using US Census Bureau data, we confirmed our prediction that the difference in compactness between the computed and actual districts would be greatest for states that are large and therefore difficult for humans to properly district given their limited capacities. The computed solutions highlighted differences in how humans and machines solve this task with machine solutions more fully optimized and displaying emergent properties not evident in human solutions. These results suggest a division of labour in which humans debate and formulate districting criteria whereas machines optimize the criteria to draw the district boundaries. We discuss how criteria can be expanded beyond notions of compactness to include other factors, such as respecting municipal boundaries, historic communities, relevant legislation, etc.

Computers and Society,Physics and Society

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is **the threat posed by partisan gerrymandering to the democratic system**. Specifically, the authors focus on how to use computational models to automate the process of district - delimitation in order to optimize objective and public criteria, such as the spatial compactness of districts. The core of the paper lies in proposing a new algorithm - the weighted k - means algorithm, and verifying the effectiveness of this algorithm through data from the US Census Bureau.

### Problems mainly solved in the paper:
1. **The threat of partisan gerrymandering**: Partisan gerrymandering refers to redrawing district boundaries to promote the interests of the controlling political faction. This practice may lead to complex and unfair district shapes, thus harming the democratic system.
2. **Limitations of human cognitive ability**: For large states, the complexity of the district - delimitation task may exceed human cognitive ability, making it difficult for humans to draw the optimal district maps.
3. **Automated solutions**: Propose an automated district - delimitation method based on computational models to reduce human intervention and potential biases and improve the fairness and transparency of district - delimitation.

### Specific methods:
- **Weighted k - means clustering algorithm**: This algorithm maximizes the compactness of districts by minimizing the average distance between voters within a district. At the same time, the algorithm also takes into account the constraint that the population of each district should be approximately equal.
- **Data sources**: Use data from the US Census Bureau, especially data from 2010 and 2015, to estimate the population distribution of each district.
- **Evaluation criteria**: Evaluate the effectiveness of the algorithm by comparing the compactness differences between the calculated district maps and the actual district maps.

### Main findings:
- **Greater improvement in large states**: For large states, the difference in compactness between the calculated district maps and the actual district maps is greater, indicating that the district - delimitation task in large states is more difficult for humans.
- **Advantages of machine solutions**: Machine - generated district maps are more optimized and show some characteristics that do not appear in human - generated solutions, such as nested district structures.
- **Public preference**: Through an online survey, most respondents prefer the computationally - generated district maps to the existing actual district maps.

### Conclusion:
The paper recommends a human - machine collaborative approach to district - delimitation, that is, humans decide and formulate the criteria for district - delimitation, and computers are responsible for optimizing district boundaries according to these criteria. This method can improve the fairness and transparency of district - delimitation and reduce the impact of human biases.