TL;DR
This paper explores whether the complexity of legal codes in US cities scales with population size, akin to other urban phenomena like GDP or infrastructure. They found that legal code volume scales sublinearly with population, suggesting it functions more like infrastructure than a direct measure of economic output, and this scaling is consistent with the increasing need to regulate social interactions in larger, denser populations.
What Problem Does This Paper Address?
The paper tackles the question of how legal systems evolve and scale with societal growth, specifically focusing on the complexity of municipal laws in US cities. While previous research in "physics of cities" has established scaling laws for urban productivity (superlinear scaling, ~1.2) and infrastructure (sublinear scaling, ~0.8), the scaling of legal complexity remained an open question. Understanding this scaling is crucial because legal codes are fundamental infrastructure that provides certainty and predictability, fostering investment and innovation. The authors aim to determine if legal code complexity grows proportionally, faster, or slower than population, and what this implies about the function of law in society.
Methods
The researchers analyzed a corpus of municipal codes from 3259 US cities. They measured "legal complexity" using several text-based metrics, primarily focusing on the file size (in bytes) of the compressed municipal codes. This approach is rooted in information theory, treating the legal code as a form of data. They also employed other measures like word count, sentence count, characters per word, words per sentence, and text entropy to assess linguistic sophistication. The core analytical method involved fitting power-law relationships (Y = aN^β) between these complexity measures and city population, using log-log regression to estimate the scaling exponent (β) and the coefficient of determination (R²). They also conducted analyses on state legislation, judicial opinions, and collective bargaining agreements to test the robustness of their findings across different legal corpora and jurisdictions.
Datasets
The primary dataset consists of municipal codes from 3259 US cities, obtained from LexisNexis. This data was pre-processed to include only ordinance texts and to de-duplicate entries. City population data was sourced from the US Census. For supporting analyses, the authors utilized:
- Texts of newly enacted state legislation and judicial opinions from US states (collected by Ash et al. [10] and Ash & MacLeod [14]).
- Corpora of collective bargaining agreements from Canada (Ash et al. [11]) and Brazil (Lagos [12]), matched with the number of covered employees as a proxy for population.
Key Insights and Ideas
- Law as Infrastructure: The sublinear scaling of legal code volume with population suggests that law, like physical infrastructure (roads, utilities), grows at a slower rate than population. This contrasts with productivity measures (GDP, patents) which scale superlinearly.
- Regulation of Social Interactions: The observed scaling parameter (β ≈ 0.2) is similar to that of GDP per capita. This suggests that legal code complexity scales with the number of social interactions per capita, reflecting the increasing need to regulate interactions between strangers as cities grow.
- Efficiency of Legal Code: The compressed and uncompressed file sizes scale at nearly identical rates. This implies that legal code does not become significantly more or less textually efficient as it grows, and the observed scaling is not due to boilerplate or redundancies.
- Linguistic Sophistication vs. Volume: While measures of linguistic sophistication (word length, sentence length, vocabulary diversity) also increase with population, their scaling exponents are much smaller and less stable than those for code volume. This indicates that increased legal complexity is primarily driven by the volume of regulation, not necessarily by more intricate language.
- Cross-Jurisdictional Consistency: The sublinear scaling pattern holds across various contexts, including US cities, state legislatures, state courts, and even private legal ordering (union contracts in Canada and Brazil), suggesting a general mechanism at play.
Key Results
- Municipal Codes: Legal complexity (measured by compressed file size) scales with city population with a scaling parameter (β) of approximately 0.2, and an R² of approximately 0.2. This relationship is linear on a log-log scale, indicating a power-law relationship.
- State Legislation: Newly enacted statutes scale with state population at a rate of β ≈ 0.4 (R² = 0.26).
- Judicial Opinions: Judicial opinions scale with state population at a higher rate of β ≈ 0.7 (R² = 0.66).
- Collective Bargaining Agreements: Canadian agreements scale with β ≈ 0.14 (R² = 0.07), and Brazilian agreements with β ≈ 0.07 (R² = 0.05).
- Linguistic Sophistication: Measures like word length, sentence length, and text entropy show positive but weak scaling relationships with population (e.g., word length β ≈ 0.003, R² = 0.08).
Limitations and Future Work
The authors acknowledge several limitations and suggest avenues for future research:
- Mechanisms: While the paper establishes the scaling relationship, the precise mechanisms driving it (e.g., volume/diversity of conflicts, economic activity) remain open questions.
- Alternative Measures: The paper notes that current complexity measures might not fully capture the "optimal encoding of types of conflicts." Future work could explore more sophisticated measures, potentially drawing from information theory, to assess legal "codewords" and their relationship to social conflict.
- Structure and Cross-References: Future research could investigate the complexity of legal codes by examining their structure (number of sections, depth of clauses) and the density of cross-references, which could reveal more about specialization and topic branching.
- Causality: The current study focuses on correlation. Establishing causal links between population growth and legal complexity would require more advanced causal research designs.
- Contextual Differences: The scaling parameters vary across different legal corpora (cities, states, courts, contracts). Understanding the reasons for these differences, such as why court opinions might be more sensitive to population, is a key area for future investigation.
Summary updated