Demographic Transition Model: Stages and Population Pyramids
A rigorous guide to the demographic transition model, its four- and five-stage versions, changing population-pyramid shapes, and the model’s limits.
The demographic transition model (DTM) summarizes a recurring historical shift from high birth and death rates to low birth and death rates. Population pyramids show the age-structure consequences of that shift: first a large surviving child population, later a narrowing base, and eventually a larger older population.
The DTM is a descriptive framework, not a universal law. It does not supply a precise timetable, explain every cause, rank countries by development, or generate a population projection on its own.
Four stages or five?
Both versions are widely taught. The classic four-stage model ends with low birth and death rates. Many textbooks add a fifth stage for prolonged very low fertility, ageing, and possible natural decrease. Some propose additional stages or reversals.
Because definitions vary, a careful analysis should describe the underlying rates rather than assert a stage number without qualification.
| Stage | Birth rate | Death rate | Natural increase | Common age-structure tendency |
|---|---|---|---|---|
| 1 | High and variable | High and variable | Low on average, volatile | Many children, rapid taper with age |
| 2 | Remains high | Falls substantially | Rapid | Very broad base and expanding young cohorts |
| 3 | Falls | Low or falling slowly | Positive but slowing | Base begins to narrow; working-age share often rises |
| 4 | Low | Low | Low, zero, or modestly positive | More even lower and middle cohorts |
| 5 | Very low in the common extension | Low | Often negative | Narrow base and larger middle/older cohorts |
Natural increase means births minus deaths. Total population growth also includes net international migration, so a population can have natural decrease and still grow.
Stage 1: high birth and death rates
In the model’s first stage, birth rates and death rates are both high. Population growth is slow on average but can be highly variable because mortality crises, harvest failures, epidemics, and conflict produce sharp fluctuations.
The idealized pyramid has many children and narrows quickly with age because relatively few people survive to older ages. However, historical age distributions may be irregular, and reliable national age-sex data are often unavailable for populations associated with this pre-transition condition.
No country today is a clean national example of textbook stage 1. Using a contemporary country as a stand-in is both empirically weak and easily turns the model into a simplistic development ranking.
Stage 2: mortality falls before fertility
In stage 2, death rates fall while birth rates remain high. Improvements can include safer water, sanitation, food security, vaccination, public health, and medical care. The exact combination and sequence vary.
More infants and children survive, so young cohorts become very large. The pyramid is usually strongly expansive, with a broad base and a steep taper. Natural increase accelerates because the gap between births and deaths widens.
The age structure also creates momentum. Large child cohorts later enter reproductive ages, sustaining many births even if fertility begins to fall. Stage 2 is therefore not merely “high fertility”; the mortality decline and the widening gap between rates are central.
Mexico in 1950 and 1975 follows this pattern closely in the UN estimates. Life expectancy rose from about 44 years in 1950 to about 59 in 1970, while the total fertility rate stayed above 6.5 births per woman and annual growth rose above 3% in the 1960s. The share under 15 rose from about 44% in 1950 to about 47% in 1975 as more children survived.
Stage 3: fertility falls and growth slows
During stage 3, birth rates decline while death rates remain comparatively low. Explanations differ among places and periods, but commonly studied factors include child survival, education, urbanization, access to contraception, the costs and benefits of childbearing, women’s opportunities, later partnership and birth timing, and changing family preferences.
The youngest bars begin to narrow relative to children and young adults just above them. This change first appears at the base; it takes decades to reshape the middle and top.
By 2000, Mexico’s fertility rate had fallen to about 2.7 births per woman. The three child bands were almost equal in width, each holding about 11–12% of the population, while the bands above them still narrowed with age.
The working-age share may rise as earlier large child cohorts mature while new child cohorts become smaller. This can create a demographic window, but an economic dividend is not automatic. Health, education, employment, gender equality, institutions, and the capacity to save and invest determine whether the age structure becomes an economic advantage.
Stage 4: low birth and death rates
With birth and death rates both low, natural increase becomes small. Cohort widths can become more even through younger and middle ages, giving a roughly stationary or column-like shape.
But the visual label can mislead. A country may still grow because of migration, improving survival, or momentum. Past baby booms and busts remain visible. Births and deaths can also be low without being equal.
This is why stage 4 should be established from vital rates and population accounting, not inferred from a rectangular outline alone.
Mexico’s projected fertility is 1.85 births per woman in 2026, and its population is still growing, at about 0.8% a year. By 2050, every five-year band from 0–4 through 55–59 holds between about 5.6% and 7.1% of the population, while growth slows to about 0.2%.
Stage 5: very low fertility, ageing, and possible natural decrease
In the common five-stage extension, fertility remains below the level required for long-run replacement in the absence of migration. Small birth cohorts narrow the base while large adult cohorts age upward. Deaths may exceed births, producing natural decrease.
The typical shape is constrictive, but total decline is not inevitable. Net immigration can offset natural decrease, and the timing depends on the inherited age structure.
In the medium projection, Mexico’s population peaks in the late 2050s and then declines. By 2075 the 0–4 band is the smallest below age 80, the 60–64 band is the largest, and about 26% of the population is 65 or older. Japan, where deaths already exceed births, shows a similar shape today.
Stage 5 is not part of every formulation, and its meaning is less standardized than the first four stages. Some accounts speculate that fertility will recover; others define the stage simply by natural decrease. State which definition you are using.
Why the pyramid changes later than the rates
A population pyramid is a record of many past birth, death, and migration regimes. When fertility drops in a given year, only the newest cohort changes immediately. Everyone already alive remains in the distribution and moves upward with time.
This lag produces several important effects:
- fertility may be below replacement while total population continues growing;
- the child share can fall while the number of children still rises;
- a working-age bulge can persist decades after fertility decline;
- ageing can accelerate when a large cohort crosses the 65 threshold;
- returning fertility to replacement would not instantly reverse ageing.
The DTM describes changes in rates. The pyramid displays the accumulated age-structure response.
Birth rate, fertility, and growth are not synonyms
These measures are often confused:
- Crude birth rate: births during a year per 1,000 population.
- Total fertility rate: births a hypothetical woman would have under the year’s age-specific fertility rates.
- Natural increase: births minus deaths.
- Population growth: natural increase plus net migration.
A young population can have a high crude birth rate even after fertility has fallen because many people are in childbearing ages. A destination can grow with low fertility because migration is positive. A population can have below-replacement fertility for years before deaths exceed births.
Why a pyramid alone cannot assign a stage
The same broad shape can result from different processes. Consider four cases:
- A broad base produced by high fertility and falling child mortality.
- A wide young-adult middle produced by labor migration, as in the United Arab Emirates.
- A narrow base produced by fertility decline.
- A missing cohort produced by war, displacement, or a past birth shock.
Only the first and third map neatly onto the simplified transition sequence. The others can mimic parts of it.
To assess a stage, examine at minimum:
- birth and death rates over time;
- fertility and life expectancy trends;
- natural increase separately from net migration;
- age-specific population change;
- whether national averages conceal major regional differences;
- data quality and boundary changes.
The model’s major limitations
It generalizes from historical experience
The model emerged from observations of populations that underwent transition under particular historical conditions. Later transitions have often occurred faster and under different medical, political, economic, and international circumstances.
It describes more readily than it explains
Saying that fertility declines in stage 3 does not identify one universal cause. Multiple mechanisms interact, and their importance differs across societies.
Migration sits awkwardly in the framework
DTM tables usually emphasize births and deaths. Yet migration can dominate growth and radically reshape working ages and sex balance, especially in small populations and labor destinations.
Stages imply sharper boundaries than the data contain
Birth and death rates move continuously, sometimes reverse, and can respond to crises. A country can sit between textbook stages or contain regions that follow different paths.
It is not a projection method
Demographic projections require explicit age-specific assumptions and cohort-component accounting. The DTM does not contain enough information to calculate future fertility, mortality, migration, or population. Published demographic research has made this logical limitation explicit.
It can invite value judgments
“Early” and “late” stages must not be treated as moral rankings. The model describes selected demographic rates, not the worth, welfare, culture, or inevitable future of a population.
DTM versus population projection
The distinction is fundamental:
- The DTM is an organizing narrative about long-run changes in birth and death rates.
- A cohort-component projection advances populations by age and sex while applying assumptions about fertility, mortality, and migration.
UN World Population Prospects uses probabilistic and demographic models informed by country data; it does not obtain a 2050 total by assigning a country to “stage 3” or “stage 4.” When you use this site’s future charts, you are viewing the UN medium scenario, not a DTM forecast.
A better classroom analysis
When given a pyramid and asked about demographic transition, use this sequence:
- Describe the evidence: compare the base, middle, top, and male/female sides.
- Measure the structure: cite the under-15 share, 65+ share, and median age.
- Inspect rates: compare fertility, growth, and—where available—birth and death rates.
- Check migration: ask whether a working-age or sex-specific bulge suggests migration.
- Compare time: see how the same cohorts move between historical and projected years.
- Assign cautiously: say “consistent with” a stage and state which version of the model you are using.
- Name uncertainty: separate historical estimates from future projections.
This approach uses the model as a hypothesis framework rather than forcing the data into a label.
Worked reasoning example
Suppose a pyramid has a broad base, 42% of the population under 15, a rapidly rising population, and a falling—but still high—fertility rate.
A careful answer would be:
The age structure is expansive and contains large child cohorts. Falling fertility suggests that the base may begin narrowing, while the large young population creates momentum. This combination is consistent with movement from the model’s second toward third stage, but birth, death, and migration series are required to support the classification.
That answer is stronger than “stage 2” because it separates what the chart shows from what the model proposes.
Sources and further reading
- United Nations World Population Prospects 2024 — demographic estimates and projections used throughout this site.
- WPP 2024 methodology report — cohort-component methods and probabilistic assumptions.
- United Nations World Social Report 2023 — population age structures across the demographic transition.
- Loschky and Wilcox, “Demographic transition: a forcing model” — formal critique of using transition theory as a projection engine.
- U.S. Census Bureau population-pyramid activity — calculating and interpreting age-sex shares.
Frequently asked questions
What are the five stages of the demographic transition model?
A common five-stage version moves from high birth and death rates, through mortality decline and then fertility decline, to low rates and, in stage 5, possible natural decrease. The original classroom model often has four stages, so stage numbering is not universal.
How does a population pyramid change through the demographic transition?
It commonly shifts from a broad triangular base toward a more column-like structure and later a narrow base with larger older cohorts. The age structure changes with a lag because existing cohorts must age through the pyramid.
Can a population pyramid identify a country’s DTM stage?
Not reliably on its own. A pyramid shows age and sex composition, while DTM stages are defined using birth and death rates. Migration, shocks, and within-country differences can produce shapes that do not fit the textbook sequence.