Fertility Stalling

Definition

Fertility stalling is the empirical phenomenon in which a country's total fertility rate (TFR) ceases to decline during the middle of a demographic transition, plateauing above replacement (TFR $> 2.1$ ) for an extended period — often decades — before either resuming decline or remaining at a higher-than-replacement steady state. Stalling is distinct from pretransitional natural fertility (which precedes any voluntary limitation) and from below-replacement equilibrium (which follows a completed transition). It represents an intermediate, durable non-convergence outcome within ongoing transitions.

Key Ideas

Empirical baseline. Bongaarts (2002) identifies Argentina and Uruguay as canonical historical examples: both entered the transition in the first half of the 20th century, declined to $\approx 3$ by the 1950s, and changed very little thereafter (TFR still above $2.5$ in 1995–2000). A Demographic and Health Surveys (DHS)-based analysis of 9 developing countries with two surveys in the 1990s found five with pace $\leq 0.03$ births/woman/year — essentially stalled given measurement uncertainty.
Development threshold. Reaching near-replacement fertility requires, on average, life expectancy $\approx 75$ years combined with adult literacy $\approx 95\%$ . In 1995–2000 only 20 of 137 developing countries met both thresholds, and their average TFR was $2.24$ — near but not at replacement. Countries far from these thresholds are candidates for prolonged stalling above replacement.
UN projection bias. Standard United Nations (UN) projection methodology interpolates linearly between a current estimate and a target year at which TFR is assumed to reach $2.1$ . This approach systematically over-projects pace for countries that stall and under-projects countries that overshoot below replacement. The errors partially offset in the aggregate but create systematic country-level errors.
Late-transition determinants. Once diffusion and social interaction effects from the early transition have run their course, fertility becomes closely tied to human development — specifically life expectancy and adult literacy. Gross domestic product (GDP) per capita, percent urban, and percent of labor force in agriculture are statistically insignificant after controlling for these two variables. Stalling therefore tends to occur when human development progress is slow.
Geographic concentration. Post-2000 research (not in Bongaarts 2002) has documented sub-Saharan Africa as the primary locus of ongoing stalling. TFR declines in the region have been slow and uneven, consistent with the development-threshold framework — most sub-Saharan countries fall well below the life expectancy and literacy thresholds associated with near-replacement fertility.
Stalling vs. wanted fertility above replacement. Stalling can reflect either (a) genuine desired family size above replacement (high wanted TFR) or (b) high unwanted fertility due to unmet need for family planning. The two mechanisms have different policy implications: the first requires addressing demand (child costs, women's education, child survival), the second addresses supply (contraceptive access, programs). Bongaarts (2002) finds that high-quality family planning programs can move countries closer to replacement than development levels alone would predict (example: Bangladesh, TFR $3.3$ despite low development scores).

How It Works

The three-phase causal structure of the fertility transition (Bongaarts 2002) explains why stalling is more likely in the middle and late phases:

Pretransition — natural fertility; TFR is high and relatively unresponsive to development. No stalling risk because decline hasn't started.
Early transition — diffusion and social interaction processes release pent-up demand for fertility limitation faster than development alone would predict. Decline is rapid. A country cannot easily "stall" here because social dynamics are self-reinforcing.
Mid/late transition — social interaction effects exhaust. Further decline requires further development (rising life expectancy and literacy). If development stalls — due to poverty traps, political instability, or slow human capital accumulation — fertility stalls with it. This is the phase where stalling occurs.

The shift from diffusion-driven to development-driven dynamics means that momentum from the early transition does not automatically carry a country through to replacement. A country that reached TFR $3.5$ through diffusion-driven decline may find that the remaining $1.5$ births require human development gains that take decades to achieve.

Why It Matters

Population projections. Stalling extends the period of positive population growth in high-fertility countries and shifts the trajectory of global population. Systematic under-prediction of stalling by UN median-variant projections leads to underestimation of future population in sub-Saharan Africa.
Development planning. The development-threshold finding implies that broad human development investments (health, education) are more effective than income growth or urbanization alone for driving late-transition fertility decline. The finding that GDP per capita is insignificant in the regression after controlling for life expectancy and literacy has direct implications for development strategy.
Family planning programs. The unwanted-fertility mechanism (unmet need) makes family planning programs especially relevant in stalling contexts. Evidence that programs move countries below their development-predicted TFR supports maintaining program infrastructure even after early-transition declines seem complete.

Open Questions

Is stalling a stable long-run equilibrium or a prolonged plateau that eventually resolves? The Argentina/Uruguay examples span 50+ years but show slow recent declines, suggesting eventual resolution is possible.
What is the relative contribution of wanted vs. unwanted fertility to stalling? Cross-country DHS decompositions exist but are sensitive to measurement of "wanted" births.
Does sub-Saharan Africa's stalling reflect the development-threshold mechanism, or does Human Immunodeficiency Virus/Acquired Immunodeficiency Syndrome (HIV/AIDS), conflict, and demographic age structure create distinct pathways?
As the life expectancy and literacy thresholds are themselves changing (with continued improvement in both globally), does the estimated replacement-threshold relationship remain stable?

Sources

Bongaarts 2002 — The End of the Fertility Transition in the Developing World