Cities can benefit from complex supply chains

Cities are heavily reliant on complex supply chain networks for their functioning. The increasing complexity of these networks, driven by globalization and lean management practices, raises concerns about vulnerability to disruptions. While complexity is often associated with instability, this research explores the counterintuitive hypothesis that supply chain complexity, specifically supplier diversity, can actually enhance a city's resilience to shocks. The study aims to understand how the architectural characteristics of urban supply networks (horizontal and vertical complexity) vary across cities and how this architecture relates to a city's ability to withstand supply chain disruptions. Measuring supply chain complexity presents challenges due to limited visibility of upstream suppliers. This paper overcomes this challenge by developing a network-based index using data on a city's immediate suppliers, alongside information on product technological sophistication, assuming that more complex products necessitate more complex upstream supply chains. This data-driven measure considers both network structure and product sophistication, key factors influencing supply chain complexity. The study uses a large dataset of annual supply flows to US cities to analyze this complex relationship, providing insights crucial for anticipating and managing supply chain risks in urban settings.

Existing literature highlights the perceived negative association between supply chain complexity and resilience to disruptions at the company level. However, the research lacks a thorough understanding of this relationship at the city level and the influence of network architecture. Ecological theory, which suggests that ecosystem stability is linked to complexity, provides a counterpoint, indicating that diversity within complexity may promote stability. The authors draw an analogy between ecological and urban supply networks, proposing that supplier diversity within a city's supply chain network could contribute to enhanced resilience. Previous studies investigating urban scaling primarily focus on internal city characteristics without explicitly considering network interactions with other cities and regions. This research addresses this gap by focusing on the network architecture of urban supply chains.

The study employs a large dataset comprising over 1 million annual supply flows to 69 major US cities from 2012 to 2015. This data is obtained from the Freight Analysis Framework version 4 (FAF4) database. The dataset covers 39 product categories, representing the US product economy. Two network measures of supply chain complexity are calculated: vertical complexity (SCI), representing supply chain depth based on the shares of a city's supply inflows, and horizontal complexity (SCI'), representing supplier diversity based on the shares of a city's number of supplier connections. The methodology uses a location quotient (LQ) equation to create a product-region bipartite network from individual product supply networks. The LQ threshold of 1 highlights key industries differentiating city supply chain structures. A dimensionality reduction algorithm, similar to the economic complexity index method, ranks the complexity of product categories based on technological sophistication. Supply chain shock intensity is calculated for each product-region pair as the largest negative annual inflow deviation from the average inflow during the study period. Nested cross-sectional regression models are used to assess the relationship between supply chain shock intensity and the complexity indices, controlling for various city and network characteristics such as population, gross metropolitan product (GMP), economic complexity index (ECI), and average shipment distance. Three sets of nested regression models are implemented to test the robustness of the findings, using different response variables (average shock intensity of main imports, average shock intensity of all imports, and individual product-region pair shock intensity) and incorporating various control variables (population, GMP, ECI, distance, percent of foreign-sourced supplies, percent of urban-sourced supplies, total product production). Additional robustness checks are conducted to account for potential limitations in the shock intensity measure.

The study reveals a consistent trade-off pattern between horizontal and vertical supply chain complexity across cities. Vertical complexity (SCI) tends to increase as horizontal complexity (SCI') decreases, indicating a contrasting relationship between supply chain depth and supplier diversity. This pattern is observed across various city characteristics (population, GMP, population density, ECI) and supply network characteristics (in-degree, out-degree, in-strength, out-strength). Regression analysis strongly demonstrates that horizontal supply chain complexity (SCI') is significantly and negatively associated with supply chain shock intensity. This negative association persists across various model specifications and sample sizes. This means cities with higher horizontal complexity tend to experience less intense shocks, on average. Specifically, a one standard deviation increase in SCI' reduces the average shock intensity by approximately 13%. While the local effect of city size (population) also shows a negative correlation with shock intensity, the network effect (SCI') is more pronounced in medium-sized cities. Large cities may experience less intense shocks due to their reduced dependence on external supply networks. The analyses suggest that improvements in horizontal supply chain complexity, especially for high-complexity products, can significantly contribute to urban resilience. The explained variance in shock intensity is higher when considering average shock intensity across multiple products rather than individual products, suggesting that this approach is more effective at predicting a city's overall shock resilience. Robustness checks, including using stationary data and average fluctuation of supply inflows as alternative response variables, consistently support the main finding.

The findings address the research question by demonstrating that contrary to common assumptions, increasing horizontal supply chain complexity, particularly supplier diversity, enhances a city's resilience to supply chain shocks. This is crucial for urban sustainability and risk management. The observed trade-off between horizontal and vertical complexity suggests that a balanced approach is necessary. The significant negative association between horizontal complexity and shock intensity provides evidence for the importance of diversifying suppliers, especially for complex products. The robustness of the findings across different model specifications and response variables reinforces the reliability of the results. The strong influence of supplier diversity (horizontal complexity) suggests that policies aimed at fostering greater diversification, particularly in high-complexity product sectors, could play a vital role in enhancing urban supply chain resilience. The complementary roles of local city effects (population) and network effects (SCI') in reducing shock intensity align with urban economic theory, where equilibrium and stability are maintained by offsetting benefits and costs. High shock intensities increase the costs to cities, emphasizing the stabilizing impact of both local and network effects.

This research contributes significantly to the understanding of urban supply chain resilience by highlighting the importance of horizontal supply chain complexity (supplier diversity) in mitigating supply chain shocks. The findings underscore the need for a holistic, multisector approach to supply chain design and policy to enhance resilience. Future research could explore the dynamics of supply chain complexity over longer time periods and investigate potential policy interventions to promote supplier diversity and improve supply chain resilience at various scales beyond cities. The data-driven approach used in this study, combined with the increasing availability of data from smart technologies, holds promise for broader applications across various supply chain actors.

The study's reliance on cross-sectional data from a relatively short time period (2012–2015) limits the ability to establish causality definitively. Future research employing longitudinal data would strengthen the causal inferences. The use of a relatively coarse level of product aggregation might mask finer-grained relationships within specific product sectors. The data's focus on the US context limits the generalizability of findings to other countries with potentially different supply chain structures and economic contexts. While the shock measure utilized is generally applicable and appropriate robustness checks were performed, future research could explore alternative metrics to capture the nuances of supply chain disruptions.