Theoretical Foundations: From Classical Transition Theory to Contemporary Multi-Process Models

The Demographic Transition Model (DTM), despite its foundational status in demography, increasingly reveals its inadequacy in capturing the intricate, interdependent dynamics that characterize contemporary population change. Conceived in the early-to-mid 20th century to describe the historical shift from high to low fertility and mortality in Western Europe and North America, the DTM posited a largely sequential, autonomous process where mortality decline preceded and subsequently triggered fertility decline, with migration treated as a residual or exogenous factor. This paradigm, while instrumental in explaining past trends, falters when applied to the simultaneous, nonlinear, and mutually reinforcing processes of aging, migration, and urbanization that define the 21st century. The classic model’s assumption of a closed population, its stage-based linearity, and its neglect of spatial economic differentials fail to account for the feedback loops where aging populations actively shape migration flows, which in turn accelerate urbanization and alter the demographic structure of both sending and receiving regions. Consequently, a theoretical retrofit is not merely an academic exercise but a methodological imperative for accurate forecasting and effective policy design. This chapter critically traces this evolution from the deterministic DTM to contemporary multi-process models that explicitly integrate the triad, establishing the analytical bedrock upon which the remainder of the book builds. It argues that only by modeling aging, migration, and urbanization as a coupled system—where changes in one component reverberate through the others—can we grasp the systemic transformations reshaping global systems.

The Classical Demographic Transition Model: Structure, Successes, and Inherent Limitations

The classic DTM, formalized by Warren Thompson (1929) and later refined by Frank Notestein (1945), delineates population change into four or five distinct stages. Stage 1 (pre-transition) is characterized by high and fluctuating birth and death rates, resulting in minimal population growth. Stage 2 (early transition) sees a sharp decline in mortality due to improvements in public health, sanitation, and food supply, while fertility remains high, leading to rapid population growth. Stage 3 (late transition) witnesses a deliberate decline in fertility driven by urbanization, industrialization, and changing socioeconomic values, causing growth rates to slow. Stage 4 (post-transition) features low, stable birth and death rates, with growth approaching zero. Some formulations add Stage 5, where fertility falls below replacement level, leading to natural decrease and population aging. The model’s explanatory power for historical Western experience is undeniable; for instance, England’s death rate began its sustained fall around 1750, while its birth rate did not decline significantly until the 1870s, creating a century of explosive growth.

However, the DTM’s limitations are profound when applied globally and contemporaneously. First, it assumes a closed system, explicitly ignoring migration. For nations with high immigration, such as the United States or Gulf Cooperation Council states, net migration contributes substantially to population growth and structural change, decoupling their trajectories from the fertility-mortality nexus predicted by the DTM. The U.S., for example, would have reached zero population growth decades earlier without sustained immigration, which now accounts for over half of its annual population increase. Second, the model implies a unidirectional, sequential causality: mortality decline *causes* fertility decline. Contemporary evidence shows a more complex interplay; in many developing countries, fertility decline began *before* mortality stabilized, influenced by family planning programs, female education, and cultural diffusion. Third, the DTM treats stages as uniform and inevitable, overlooking path dependencies and institutional contexts. Sub-Saharan Africa, for instance, exhibits stalled transitions where fertility remains high despite significant mortality declines, challenging the model’s teleology. Fourth, and most critically for this book, the DTM is aspatial. It does not incorporate the spatial redistribution of populations—urbanization—or the economic differentials that drive migration, both of which are now primary engines of demographic change. The model cannot explain why a country like Brazil achieved sub-replacement fertility while Nigeria did not, a divergence heavily mediated by differential urbanization patterns and internal migration flows.

Toward Multi-Process Models: Integrating Fertility, Mortality, and Migration

The recognition of these limitations catalyzed the development of multi-process demographic models that explicitly simulate interactions among fertility, mortality, and migration. These models reject the notion of demographic components as independent streams, instead treating them as interdependent processes influenced by common socioeconomic determinants and by each other.

A seminal contribution is the unified demographic framework developed by demographers such as Samuel H. Preston and Ansley Coale, which employs structural equation modeling to parse the direct and indirect effects of development indicators (e.g., GDP per capita, female education) on demographic outcomes. For example, analysis of Demographic and Health Survey (DHS) data across 70 countries reveals that female secondary school enrollment reduces fertility both directly (by changing preferences) and indirectly (by improving child survival, which in turn lowers desired family size). This fertility-mortality linkage is a core component of the multi-process approach.

Migration is integrated not as an afterthought but as a systemic response to spatial economic differentials. The Harris-Todaro model (1970), originally an economic model of rural-urban migration in developing countries, provides a behavioral foundation: individuals migrate if the *expected* urban income exceeds the rural income, where expected income accounts for the probability of urban unemployment. Formally: $$ E[W_{urban}] = W_{urban} \times (1 - u_{urban}) > W_{rural} $$ where $W$ represents wage and $u$ unemployment. This model predicts that even with high urban unemployment, migration persists because the potential upside outweighs rural stagnation. Empirically, it explains the massive rural-to-urban flows in China during the 1980s-2000s, where despite urban unemployment estimates of 5-10%, the expected income gain was substantial. The model has been extended to international migration by incorporating migration costs, network effects, and destination country policies, as seen in the migration-specific capital model where prior migrants reduce the risks and costs for subsequent ones.

At the operational level, the cohort-component method remains the standard projection framework but is now routinely enhanced to incorporate multi-process interactions. The basic formula for projecting a population by age and sex is: $$ P_{t+n}(x+n) = P_t(x) \times S(x) + B_t \times \text{sex ratio at birth} + M_{in}(x) - M_{out}(x) $$ where $S(x)$ is the age-specific survival probability, $B_t$ births, and $M$ net migration by age. Advanced implementations, such as those by the U.S. Census Bureau or the European Commission’s AGEING-ECON project, model $S(x)$ as a function of health behaviors influenced by migration (e.g., the “healthy migrant effect” where immigrants have lower mortality than natives), and model fertility $B_t$ as a function of female labor force participation, which itself is affected by migration (e.g., immigrant women often have higher fertility). These models use microsimulation or dynamic stochastic general equilibrium (DSGE) approaches to capture feedbacks. For instanc