Migration Models (Push-Pull / Multiregional)
Also known as: Push-Pull Migration Theory, Multiregional Migration Model, Lee Migration Framework, Göç Modelleri
Migration models are quantitative frameworks for explaining and forecasting population movement between geographic units. Lee's (1966) push-pull theory classifies factors at origin and destination into positive and negative forces, modulated by intervening obstacles. Widely used by demographers, regional planners, and policy researchers to project labor mobility, refugee flows, and urbanization trends across national and subnational geographies.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Apply migration models when you have origin-destination flow data and wish to decompose migration into structural determinants, project future flows, or evaluate policy scenarios. Key assumptions include rational actor behavior, measurable push-pull covariates, and stable distance-decay relationships. The framework suits internal and international migration analysis. It is less appropriate when flow data are sparse, when migration is driven by sudden unobservable shocks, or when agent-level heterogeneity is the central focus—in those cases spatial microsimulation or agent-based models may be preferable.
Strengths & limitations
- Conceptually transparent: the push-pull taxonomy is intuitive and maps directly to observable socioeconomic indicators.
- Scalable from bilateral two-region models to full multiregional systems covering hundreds of zones.
- Supports policy counterfactuals by adjusting destination attributes or obstacle parameters.
- Well-established theoretical grounding with decades of empirical validation across diverse contexts.
- Aggregate flow data mask individual-level heterogeneity; group-level inference does not apply to individual decisions.
- Distance-decay parameter β is empirically estimated and may not transfer across time periods or migration corridors.
- The model assumes that push and pull factors are observable and measurable, which is often difficult for conflict- or climate-driven displacement.
- Multiregional extensions require large, consistent origin-destination matrices that are rarely available for developing-country contexts.
Frequently asked
What data are required to calibrate a push-pull migration model?
At minimum you need origin-destination flow counts (ideally by age and sex) for one or more reference periods, plus covariates representing push and pull factors at each location—typically GDP per capita, unemployment rate, housing cost, and amenity indices. Distance or travel-time matrices between all zone pairs are also required to estimate the distance-decay parameter.
How does Lee's push-pull model differ from the gravity model of migration?
The gravity model is a mathematical specification that operationalizes Lee's conceptual framework. Lee's theory defines the categories of factors (push, pull, intervening obstacles, personal factors) without prescribing a functional form, whereas the gravity model adopts a specific multiplicative form with population masses and a distance-decay exponent. The two are complementary: Lee provides theory, gravity provides an estimable equation.
Can migration models handle non-linear or threshold effects?
Standard linear push-pull and gravity specifications cannot capture thresholds—for example, migration that accelerates sharply once unemployment crosses a critical level. Extensions using non-linear regression, regime-switching models, or machine-learning augmentation can incorporate such effects, but require richer data and careful out-of-sample validation to avoid overfitting.
Sources
- Lee, E. S. (1966). A theory of migration. Demography, 3(1), 47–57. DOI: 10.2307/2060063 ↗
How to cite this page
ScholarGate. (2026, June 2). Migration Models (Push-Pull / Multiregional). ScholarGate. https://scholargate.app/en/demography/migration-models
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Cohort-Component ProjectionDemography↔ compare
- Radiation ModelSpatial analysis↔ compare
- Spatial Interaction ModelSpatial analysis↔ compare