Skip to main content
v2026.11,986 entries · CC-BY 4.0

Economic Geography: Definition, Models, & Methods

Economic geography examines how spatial and institutional structures shape production, trade, and regional disparities using quantitative location metrics.

Written and maintained by CASRAI Editorial Board

Published · Last updated

On this page

What is economic geography? Definition and core scope

Researchers in human geography, spatial economics, and regional science study economic geography to understand how economic activity is organized across space, why industries cluster, and how spatial structures drive uneven development. The field investigates the production, distribution, exchange, and consumption of resources across geographic scales.

Economic geography is the scientific study of the spatial organization of economic systems, analyzing the location of economic activities and the geographic distribution of wealth, employment, and innovation. Framed in The New Oxford Handbook of Economic Geography (Clark et al., 2018) and the American Association of Geographers (AAG) Economic Geography Specialty Group charter, the discipline encompasses three core paradigms: (1) New Economic Geography (NEG), advanced by Paul Krugman, Masahisa Fujita, and Anthony Venables, which formalizes spatial agglomeration through models with increasing returns to scale and transport costs; (2) Evolutionary Economic Geography (EEG), led by Ron Boschma and Koen Frenken, which models regional branching and path dependency; and (3) Institutional and Relational Economic Geography, examining how social networks and governance embed economic activity in specific places. Economic geographers integrate fieldwork with spatial analytics, drawing on political ecology to evaluate resource governance, spatial mobility patterns to assess regional inequality, and workforce commuting data to delineate functional labor markets.

Quantitative formulas: Location Quotient and Krugman Specialization Index

Quantitative economic geography relies on formal metrics to evaluate regional industrial concentration and structural divergence across geographic areas. Two primary indices serve as standard analytical baselines: the Location Quotient (LQ) and the Krugman Specialization Index (KSI).

The Location Quotient, standardized by the U.S. Bureau of Labor Statistics (BLS) and regional economic base theory, measures local industry concentration relative to a benchmark reference economy:

LQ_ij = (e_ij / e_j) / (E_i / E)

Where:

  • LQ_ij: Location quotient for industry i in region j.
  • e_ij: Local employment in industry i within region j.
  • e_j: Total local employment across all industries within region j (e_j = Σ_i e_ij).
  • E_i: Reference national employment in industry i.
  • E: Total reference national employment across all industries (E = Σ_i E_i).
  • s_ij = e_ij / e_j: Local employment share of industry i.
  • s_ik = E_i / E: Benchmark national employment share of industry i.

To quantify structural divergence across the entire regional economy, economic geographers apply the Krugman Specialization Index (Krugman, 1991):

K_jk = Σ_i |s_ij – s_ik|

Where K_jk sums absolute sectoral share deviations across all N mutually exclusive industries between region j and benchmark k, ranging from 0 (identical distribution) to 2 (complete non-overlap across all industries).

Worked example: calculating regional industrial specialization

Consider an empirical assessment of the Cascadia Metropolitan Area (Region j, total employment e_j = 250,000) evaluated against the national economy (Benchmark k, total employment E = 150,000,000) across four mutually exclusive sectors:

  • Sector 1 (Biomedical Manufacturing): e_1j = 22,500; E_1 = 4,500,000.
  • Sector 2 (Information Technology): e_2j = 35,000; E_2 = 9,000,000.
  • Sector 3 (Healthcare & Education): e_3j = 50,000; E_3 = 36,000,000.
  • Sector 4 (Retail & Other Services): e_4j = 142,500; E_4 = 100,500,000.

Step 1: Compute sectoral shares (s_ij = e_ij / e_j and s_ik = E_i / E):

  • Sector 1: s_1j = 22,500 / 250,000 = 0.0900 (9.00%); s_1k = 4,500,000 / 150,000,000 = 0.0300 (3.00%).
  • Sector 2: s_2j = 35,000 / 250,000 = 0.1400 (14.00%); s_2k = 9,000,000 / 150,000,000 = 0.0600 (6.00%).
  • Sector 3: s_3j = 50,000 / 250,000 = 0.2000 (20.00%); s_3k = 36,000,000 / 150,000,000 = 0.2400 (24.00%).
  • Sector 4: s_4j = 142,500 / 250,000 = 0.5700 (57.00%); s_4k = 100,500,000 / 150,000,000 = 0.6700 (67.00%).

Step 2: Calculate Location Quotients (LQ_ij = s_ij / s_ik):

  • LQ_Biomed: 0.0900 / 0.0300 = 3.000
  • LQ_IT: 0.1400 / 0.0600 ≈ 2.333
  • LQ_Health: 0.2000 / 0.2400 ≈ 0.833
  • LQ_Retail: 0.5700 / 0.6700 ≈ 0.851

Step 3: Calculate the Krugman Specialization Index by summing absolute share deviations:

K_jk = |0.0900 – 0.0300| + |0.1400 – 0.0600| + |0.2000 – 0.2400| + |0.5700 – 0.6700| = 0.0600 + 0.0800 + 0.0400 + 0.1000 = 0.2800

Cascadia exhibits K = 0.2800, indicating moderate structural specialization driven by concentrations in biomedical manufacturing and information technology.

Interpretation and reading guidance

Interpreting spatial economic metrics requires comparing observed indices against established empirical thresholds:

Metric & Range Economic Classification Theoretical Interpretation Policy & Planning Implication
LQ ≥ 1.25 Basic (Export) Sector Local production exceeds local demand; sector brings external income into the region. Anchor sector for cluster initiatives such as regional innovation engines.
1.00 ≤ LQ < 1.25 Self-Sufficient / Emerging Local production satisfies regional demand at baseline parity. Monitor shift-share trends to identify emerging competitive advantages.
0.75 ≤ LQ < 1.00 Modest Deficit Sector Local employment is slightly below national parity; minor import reliance. Standard consumer services requiring monitoring but limited direct intervention.
LQ < 0.75 Non-Basic (Import-Dependent) Substantial regional deficit; local market relies heavily on external supply. High regional economic leakage; candidate for import-substitution strategies.
K_jk < 0.20 Diversified Economy Industrial profile closely mirrors national benchmark; high shock resilience. Broad-based infrastructure and general workforce development investments.
0.20 ≤ K_jk < 0.50 Moderate Specialization Core comparative advantage in key sectors (e.g., Cascadia: K = 0.28). Support specialized cluster infrastructure while maintaining economic agility.
K_jk ≥ 0.50 Highly Specialized Severe concentration in narrow activities (e.g., single-industry towns). Elevated exposure to structural shocks; priority for targeted diversification.

Methodological assumptions and common mistakes

Empirical spatial analysis requires rigorous accounting for geographic boundaries and aggregation mechanics. Analysts must navigate five common pitfalls:

  1. The Modifiable Areal Unit Problem (MAUP): Specialization metrics vary when spatial units are aggregated into different boundary shapes or scales. Ground boundaries in functional economic units, such as commuting zones.
  2. Uniform Productivity Assumption: The Location Quotient assumes local labor productivity equals the national baseline. If a regional industry utilizes high-capital automation, employment LQs underestimate actual output. Triangulate employment with wage earnings or regional output data.
  3. Sectoral Aggregation Bias: Evaluating specialization at broad 2-digit NAICS levels obscures fine-grained cluster dynamics. High-tech manufacturing niches can be masked inside generic manufacturing aggregates. Analysis should use 4-digit or 6-digit NAICS codes.
  4. Corporate Headquarters Distortions: Multi-establishment enterprises often report total corporate employment at administrative headquarters rather than operational plants, artificially inflating urban office LQs while undercounting regional manufacturing.
  5. Static Equilibrium Assumption: Cross-sectional LQs capture a single snapshot without reflecting temporal trajectories. A high LQ in a shrinking sector reflects industrial legacy rather than growth. Pair static LQs with dynamic shift-share models.

How to report economic geography metrics (APA example)

When reporting spatial economic metrics in scholarly manuscripts, follow American Psychological Association (APA 7th edition) guidelines, specifying geographic units, benchmarks, data sources, and statistical outcomes:

An economic base analysis of the Cascadia Metropolitan Statistical Area (total employment N = 250,000) was conducted using 2024 Quarterly Census of Employment and Wages (QCEW) data relative to the national benchmark (N = 150,000,000). The region demonstrated significant agglomeration in Biomedical Manufacturing (NAICS 3254), with local employment (e1 = 22,500, local share = 9.00%) exceeding national shares (E1 = 4,500,000, national share = 3.00%), yielding a Location Quotient LQ = 3.00. Under economic base standards (LQ ≥ 1.25), biomedical manufacturing represents a primary basic export sector. Across four mutually exclusive sectors, the region recorded a Krugman Specialization Index K = 0.28, reflecting moderate structural divergence from the national economy driven by technology and life sciences concentrations.

In tabular presentations, include raw employment figures alongside derived shares and LQs, defining all notation in table notes.

How to calculate spatial economic metrics in Excel, R, and Python

Researchers calculate Location Quotients and specialization indices across standard analytical environments using consistent logic:

Excel Formula Implementation

With local employment in cells B2:B5 and national benchmark employment in cells C2:C5:

# Location Quotient for Row 2 (Enter in cell D2 and copy down):
=(B2/SUM($B$2:$B$5))/(C2/SUM($C$2:$C$5))

# Krugman Specialization Index:
=SUM(ABS((B2:B5/SUM($B$2:$B$5)) - (C2:C5/SUM($C$2:$C$5))))

R Implementation

calc_spatial_metrics <- function(local_emp, nat_emp) {
  s_local <- local_emp / sum(local_emp)
  s_nat <- nat_emp / sum(nat_emp)
  list(lq = round(s_local / s_nat, 4), krugman = round(sum(abs(s_local - s_nat)), 4))
}
local <- c(Biomed = 22500, IT = 35000, Health = 50000, Retail = 142500)
nat <- c(Biomed = 4500000, IT = 9000000, Health = 36000000, Retail = 100500000)
calc_spatial_metrics(local, nat)

Python Implementation

import numpy as np
import pandas as pd

def compute_spatial_indices(df, local_col='local_emp', nat_col='nat_emp'):
    s_local = df[local_col] / df[local_col].sum()
    s_nat = df[nat_col] / df[nat_col].sum()
    df['lq'] = s_local / s_nat
    return df, np.sum(np.abs(s_local - s_nat))

data = pd.DataFrame({
    'sector': ['Biomedical', 'IT_Software', 'Healthcare', 'Retail_Other'],
    'local_emp': [22500, 35000, 50000, 142500],
    'nat_emp': [4500000, 9000000, 36000000, 100500000]
})
df_out, krugman = compute_spatial_indices(data)
print(df_out[['sector', 'lq']], f"Krugman Index: {krugman:.4f}")

Frequently asked questions

What is the core distinction between economic geography and regional economics?

Regional economics applies neoclassical optimization models and price theory to spatial units. Economic geography, rooted in human geography, investigates place-specific historical path dependencies, institutional governance, power relations, and geographical unevenness through mixed qualitative and quantitative methods.

What threshold indicates an export-base industry when analyzing Location Quotients?

While LQ = 1.00 reflects exact parity with national benchmarks, regional scientists apply a threshold of LQ ≥ 1.25 to identify genuine basic export industries, providing a buffer for regional consumption variances and data anomalies.

How does the Modifiable Areal Unit Problem (MAUP) impact spatial economic research?

MAUP encompasses scale effects (aggregation size) and zone effects (boundary configuration). Evaluating industrial concentration across arbitrary administrative boundaries rather than functional labor commuting zones can produce distorted cluster boundaries or dilute actual agglomeration intensity.

What is the difference between New Economic Geography (NEG) and Evolutionary Economic Geography (EEG)?

New Economic Geography (NEG) uses formal mathematical models with monopolistic competition to explain how increasing returns and transport costs generate core-periphery agglomerations. Evolutionary Economic Geography (EEG) models historical path dependency and regional branching into related technological sectors.

How do global production networks (GPNs) differ from traditional supply chains?

Traditional supply chain analysis focuses on linear logistics and cost minimization. Global Production Networks (GPN) analysis examines multi-scalar power networks, evaluating how value is created, enhanced, and captured across corporate, state, labor, and institutional actors embedded in geographic territories.

Sources

Follow CASRAI

Research-administration guidance, standards updates and independent tool reviews.

Ask CASRAI · free to try

Ask about Economic Geography: Definition, Models, & Methods

2 free questions a day, no account.

An AI assistant for research: funder rules, compliance, methods and publishing. It links the sources behind its answers and labels answers that come from a web search.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays open to read, with no account and no card.

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →