Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cities cluster into growth regimes that propagate shocks

Isaak Mengesha, Debraj Roy

Abstract

Economic growth is conventionally analyzed at the national level, yet cities generate the bulk of global output. Here we construct GDP trajectories for 8,808 functional urban areas (FUAs) across 165 countries over 1993-2019 using satellite-derived nighttime light data and identify 17 distinct, persistent growth regimes through clustering of full temporal trajectories. Rather than converging toward a common frontier, FUAs inhabit distinct economic niches-analogous to ecological niches-defined by shared volatility profiles, shock responses, and long-run dynamics that transcend national boundaries. Cities within the same country frequently belong to different regimes, while structurally similar cities on different continents share the same one; regime membership explains 16% of within-country growth variance beyond country fixed effects. National-level convergence emerges as an aggregation artifact: conditional convergence operates within regimes, not globally. A directed propagation network reveals that shocks transmit along lines of structural similarity rather than geographic proximity, with advanced economies exporting disturbances and emerging economies absorbing or amplifying them. Within-country spatial inequality declines with industrialization maturity, consistent with growth initially concentrating in leading cities before diffusing across the urban system. The global economy is better understood as an ecology of heterogeneous urban growth regimes than as a collection of nations on a shared development path.

Cities cluster into growth regimes that propagate shocks

Abstract

Economic growth is conventionally analyzed at the national level, yet cities generate the bulk of global output. Here we construct GDP trajectories for 8,808 functional urban areas (FUAs) across 165 countries over 1993-2019 using satellite-derived nighttime light data and identify 17 distinct, persistent growth regimes through clustering of full temporal trajectories. Rather than converging toward a common frontier, FUAs inhabit distinct economic niches-analogous to ecological niches-defined by shared volatility profiles, shock responses, and long-run dynamics that transcend national boundaries. Cities within the same country frequently belong to different regimes, while structurally similar cities on different continents share the same one; regime membership explains 16% of within-country growth variance beyond country fixed effects. National-level convergence emerges as an aggregation artifact: conditional convergence operates within regimes, not globally. A directed propagation network reveals that shocks transmit along lines of structural similarity rather than geographic proximity, with advanced economies exporting disturbances and emerging economies absorbing or amplifying them. Within-country spatial inequality declines with industrialization maturity, consistent with growth initially concentrating in leading cities before diffusing across the urban system. The global economy is better understood as an ecology of heterogeneous urban growth regimes than as a collection of nations on a shared development path.
Paper Structure (19 sections, 16 equations, 9 figures, 2 tables)

This paper contains 19 sections, 16 equations, 9 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Results
  3. Spatial structure and regime composition
  4. Temporal dynamics and regime trajectories
  5. Shock propagation and network structure
  6. Discussion
  7. Methods
  8. Spatial aggregation procedure
  9. Growth trajectory construction and regime clustering
  10. Preprocessing and dimensionality reduction
  11. Clustering via k-means
  12. Optimal cluster selection via silhouette score
  13. Shock detection and propagation network
  14. Lagged correlation and propagation network construction
  15. Propagation roles and network metrics
  16. ...and 4 more sections

Figures (9)

  • Figure 1: Global functional urban areas cluster into 17 distinct growth regimes transcending national boundaries. Spatial distribution of 8,808 FUAs colored by regime membership (PCA--k-means on 1993--2019 growth trajectories). Regimes exhibit geographic coherence despite no geographic variable entering the clustering: the former Soviet bloc forms a contiguous region (Cluster 6); China dominates Cluster 1; emerging Asian economies group in Cluster 0; and Latin America clusters into Cluster 2. Yet regimes also cross national borders---Mexico clusters with the US and Canada (Cluster 3)---while large countries contain FUAs from multiple regimes. Venezuela and Iraq each constitute single-country regimes driven by extreme political and resource shocks.
  • Figure 2: Each regime exhibits a distinct temporal growth signature. Mean growth trajectories $\bar{g}_{c,t}$ for each of the 17 clusters, 1993--2019. Volatility ranges from 0.66% (Cluster 10) to 18.93% (Cluster 7, Iraq). Advanced economies (Cluster 3) show synchronized response to the 2008--2009 crisis; rapidly growing Asian clusters (Clusters 0, 1) display higher baseline growth with pronounced 1997--1998 crises; resource-dependent regimes (Cluster 2) follow commodity price cycles; and conflict-affected regimes (Clusters 4, 6, 7) exhibit large discrete shocks.
  • Figure 3: Growth shocks propagate through a sparse directed network where structural similarity, not geography, determines transmission channels. Directed network of statistically significant ($p \leq 0.05$) lagged correlations between cluster growth rates, with edge width proportional to correlation strength and arrow direction indicating temporal precedence (Cluster A $\to$ Cluster B implies growth fluctuations in A precede and predict those in B with lag 0--3 years). Cluster 3 (advanced economies) acts as a net exporter (outgoing strength = 0.051), with edges pointing primarily outward to emerging-market and resource-dependent clusters. Clusters 1 (China) and 11 (West Africa) are net absorbers, receiving predominantly incoming edges. Amplifier regimes display dense bidirectional connectivity; buffer regimes (Clusters 5, 9, 10, 14, 15) are peripherally positioned with minimal edges. Note that lagged correlations establish temporal precedence but do not identify causal mechanisms; co-exposure to common upstream shocks remains a possible alternative explanation.
  • Figure 4: Shock timing and severity vary systematically across regimes. Heatmap of shock years (top and bottom 2% of each regime's growth distribution) across all 17 clusters, 1993--2019. Some shocks are globally synchronized (1997--1998 Asian crisis, 2008--2009 financial crisis, 2014--2016 commodity collapse), while others remain regime-specific (Iraq 2003--2005, Zimbabwe 2000--2008). The heterogeneity in shock timing contradicts the assumption of a common global shock process underlying standard growth regressions.
  • Figure 5: Within-country spatial inequality in FUA growth rates declines with industrialization maturity. Each point represents a country with multiple FUAs; point size is proportional to the number of FUAs. The $x$-axis shows years since the country first had industrial employment exceed agricultural employment (log scale). The $y$-axis shows the cross-sectional standard deviation of FUA growth rates, averaged over 1993--2019. The fitted line follows a log-linear specification.
  • ...and 4 more figures