shiftshare
Regional economic analysis on region x sector employment tables.
shiftshare collects the standard tools of regional economic analysis used in planning
practice and regional science, working on a matrix of employment (or output, or
establishments) by region and sector:
- Location quotients: simple (SLQ), cross-industry (CILQ), Flegg (FLQ) with a
user-chosen
deltaorlambda*, augmented FLQ (AFLQ), and regionalisation of national input-output coefficients. - Shift-share: classic (Dunn 1960), Esteban-Marquillas (1972) with homothetic employment and allocation effect, Arcelus (1984) eight-component model, and dynamic shift-share (Barff & Knight 1988).
- Specialisation and localisation: Krugman specialisation index (against the nation, the rest of the nation, or pairwise), Hoover coefficients of localisation and specialisation, localisation curve with locational Gini, Herfindahl index.
- Agglomeration: Ellison-Glaeser (1997) index and coagglomeration index, plant Herfindahl from plant sizes.
- Economic base: basic employment and base multiplier by the assumption, location quotient and minimum requirements (Ullman & Dacey 1960) methods.
It needs only NumPy. Inputs can be NumPy arrays, nested lists, or labelled
EmploymentTables read from CSV (wide or long) or JSON. A command-line tool is included.
Installation
pip install shiftshare
Requires Python 3.9 or newer.
Quickstart
import numpy as np
import shiftshare as ss
# rows = regions, columns = sectors
base = np.array([[60, 40], [40, 60]], dtype=float)
end = np.array([[75, 40], [45, 70]], dtype=float)
ss.location_quotient(base) # [[1.2, 0.8], [0.8, 1.2]]
res = ss.classic(base, end) # Dunn (1960)
res.national_share # [[9, 6], [6, 9]]
res.industry_mix # [[3, -2], [2, -3]]
res.regional_shift # [[3, -4], [-3, 4]]
res.region_totals() # each component summed per region
em = ss.esteban_marquillas(base, end)
em.competitive, em.allocation # split of the regional shift
em.allocation_type() # "competitive advantage, specialized", ...
ss.arcelus(base, end).components() # eight components that add up to the change
ss.dynamic([base, end, end * 1.05]) # year-by-year shift-share, summed
ss.krugman_index(base, versus="rest") # [0.4, 0.4]
ss.coefficient_of_localization(base) # Hoover, per sector
ss.lq_method(base).multiplier # economic base multiplier, LQ method
ss.minimum_requirements(base).multiplier # Ullman & Dacey
ss.flq(base[0], base.sum(axis=0), delta=0.3) # FLQ matrix for region 0
With labels:
t = ss.read_table("employment_2015.csv") # wide: region,Agri,Manuf,... or long
t.label(ss.location_quotient(t), "lq") # [{"region":..., "sector":..., "lq":...}]
ss.profile(t) # LQs, specialisation, localisation, multipliers
ss.decompose(t, ss.read_table("employment_2023.csv"), "esteban-marquillas")
National totals default to the column sums of the matrix (the regions make up the
nation). When the regions are only part of the nation, pass the national sector totals as
reference= (structure measures) or national_base= / national_end= (shift-share).
Features
| Area | Functions |
|---|---|
| Location quotients | location_quotient, cross_industry_lq, flegg_lambda, flq, aflq, regionalize |
| Shift-share | classic, esteban_marquillas, arcelus, dynamic, growth_rates |
| Specialisation | krugman_index, krugman_pairwise, coefficient_of_specialization, herfindahl, sector_shares |
| Localisation | coefficient_of_localization, localization_curve |
| Agglomeration | ellison_glaeser, coagglomeration, raw_concentration, plant_herfindahl |
| Economic base | assumption_method, lq_method, minimum_requirements, base_multiplier |
| Tables and I/O | EmploymentTable, read_table, read_vector, write_csv, write_json, profile, decompose, decompose_dynamic |
Methods and sources
Notation: e[r,i] employment of sector i in region r, e[r] regional total, E[i]
national sector total, E national total; 0 base year, 1 end year; G = E1/E0 - 1,
G[i] = E1[i]/E0[i] - 1, g[r] = e1[r]/e0[r] - 1, g[r,i] = e1[r,i]/e0[r,i] - 1.
| Measure | Definition | Source |
|---|---|---|
| SLQ | (e[r,i]/e[r]) / (E[i]/E) |
Isard (1960); Miller & Blair (2009) |
| CILQ | SLQ[i] / SLQ[j] (supplying i, purchasing j) |
Miller & Blair (2009) |
| FLQ | lambda* CILQ[i,j], diagonal lambda* SLQ[i], lambda* = [log2(1 + RE/NE)]^delta, 0 <= delta < 1 |
Flegg, Webber & Elliott (1995) |
| AFLQ | FLQ[i,j] log2(1 + SLQ[j]) where SLQ[j] > 1 |
Flegg & Webber (2000) |
| Regionalisation | r[i,j] = a[i,j] min(LQ, 1) |
Miller & Blair (2009) |
| Classic shift-share | NS = e0 G, IM = e0 (G[i] - G), RS = e0 (g[r,i] - G[i]) |
Dunn (1960) |
| Esteban-Marquillas | e_hat = e0[r] E0[i]/E0; C' = e_hat (g[r,i] - G[i]), A = (e0 - e_hat)(g[r,i] - G[i]) |
Esteban-Marquillas (1972) |
| Arcelus | rate split G + (G[i]-G) + (g[r]-G) + [(g[r,i]-g[r]) - (G[i]-G)], each applied to e_hat and e0 - e_hat |
Arcelus (1984) |
| Dynamic shift-share | classic model on each consecutive pair of years, summed | Barff & Knight (1988) |
| Krugman index | `sum_i | s[r,i] - s_ref[i] |
| Coefficient of specialisation | `0.5 sum_i | e[r,i]/e[r] - E[i]/E |
| Coefficient of localisation | `0.5 sum_r | e[r,i]/E[i] - e[r]/E |
| Locational Gini | 2 x area between the localisation curve (regions by decreasing LQ) and the diagonal (Krugman reports the undoubled area, maximum 0.5) |
Hoover (1936); Krugman (1991, app. D) |
| Herfindahl | sum_i s[r,i]^2 |
Hirschman (1964) |
| Ellison-Glaeser | gamma = (G - (1 - sum x^2) H) / ((1 - sum x^2)(1 - H)), G = sum_r (s_r - x_r)^2 |
Ellison & Glaeser (1997) |
| Coagglomeration | [G/(1 - sum x^2) - H - sum_j gamma_j w_j^2 (1 - H_j)] / (1 - sum_j w_j^2) |
Ellison & Glaeser (1997) |
| LQ basic employment | max(e[r,i] - e[r] E[i]/E, 0) |
Isard (1960); Tiebout (1962) |
| Minimum requirements | max(e[r,i] - m[i] e[r], 0), m[i] the smallest share among comparable regions |
Ullman & Dacey (1960) |
| Base multiplier | total / basic |
Tiebout (1962) |
Conventions and choices:
- Every shift-share decomposition adds up exactly to
e1 - e0for every cell. Cells with zero base employment have no growth rate; their change is booked as regional shift (classic), as competitive effect with zero allocation (Esteban-Marquillas, flagged inundefined), or in the regional industry-mix term (Arcelus). - No calibration constants are built in. Flegg's
delta, the plant Herfindahl for the Ellison-Glaeser index, assumption-method basic sectors or fractions, and the comparable region set and rank for minimum requirements are all supplied by the user.
References:
- Arcelus, F.J. (1984) An extension of shift-share analysis. Growth and Change 15(1), 3-8.
- Barff, R.A. and Knight, P.L. (1988) Dynamic shift-share analysis. Growth and Change 19(2), 1-10.
- Dunn, E.S. (1960) A statistical and analytical technique for regional analysis. Papers and Proceedings of the Regional Science Association 6, 97-112.
- Ellison, G. and Glaeser, E.L. (1997) Geographic concentration in U.S. manufacturing industries: a dartboard approach. Journal of Political Economy 105(5), 889-927.
- Esteban-Marquillas, J.M. (1972) A reinterpretation of shift-share analysis. Regional and Urban Economics 2(3), 249-255.
- Flegg, A.T., Webber, C.D. and Elliott, M.V. (1995) On the appropriate use of location quotients in generating regional input-output tables. Regional Studies 29(6), 547-561.
- Flegg, A.T. and Webber, C.D. (2000) Regional size, regional specialization and the FLQ formula. Regional Studies 34(6), 563-569.
- Hirschman, A.O. (1964) The paternity of an index. American Economic Review 54(5), 761-762.
- Hoover, E.M. (1936) The measurement of industrial localization. Review of Economics and Statistics 18(4), 162-171.
- Isard, W. (1960) Methods of Regional Analysis: An Introduction to Regional Science. MIT Press.
- Krugman, P. (1991) Geography and Trade. MIT Press.
- Miller, R.E. and Blair, P.D. (2009) Input-Output Analysis: Foundations and Extensions, 2nd ed. Cambridge University Press.
- Tiebout, C.M. (1962) The Community Economic Base Study. Committee for Economic Development.
- Ullman, E.L. and Dacey, M.F. (1960) The minimum requirements approach to the urban economic base. Papers and Proceedings of the Regional Science Association 6, 175-194.
Command line
shiftshare profile employment_2015.csv --csv lq.csv -o profile.json
shiftshare decompose employment_2015.csv employment_2023.csv --method esteban-marquillas \
--national-base nation_2015.csv --national-end nation_2023.csv --csv components.csv
shiftshare dynamic emp_2019.csv emp_2020.csv emp_2021.csv emp_2022.csv
shiftshare base employment_2023.csv --method mr --csv basic.csv
shiftshare base employment_2023.csv --method assumption --basic-sectors Manufacturing,Mining
shiftshare flq employment_2023.csv --region Coast --delta 0.3 --coefficients national_A.csv
shiftshare eg plants_by_region.csv --herfindahl plant_herfindahl.csv
Tables are CSV (wide: first column region, one column per sector; or long: region,
sector, employment/value columns; ,, ; or tab separated) or JSON. Sector vectors
(national totals, Herfindahl values) are CSV sector,value rows or a JSON object. Results
are printed as JSON or written with -o; --csv writes the detailed table.
Citation
If you use shiftshare in your work, please cite it (see CITATION.cff):
@software{eminoglu_shiftshare,
author = {Eminoğlu, Yusuf},
title = {shiftshare: regional economic analysis on region x sector employment tables},
version = {0.1.1},
doi = {10.5281/zenodo.24235554},
year = {2026},
url = {https://gitlab.com/geophilo1/shiftshare_sdk}
}
and the original sources of the methods you use.
License
MIT. See LICENSE.
Metadata
Release files for shiftshare 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| shiftshare-0.1.1.tar.gz | 41.2 kB | Without provenance | Details for shiftshare-0.1.1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| shiftshare-0.1.1-py3-none-any.whl | Python 3 | none | any | Details for shiftshare-0.1.1-py3-none-any.whl |
Total release size: 75.6 kB
Security for shiftshare 0.1.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.