Davis and Blake’s Model: Intermediate Determinants of Fertility

Why do some societies have more children per family than others? It’s tempting to point straight to economics, religion, or education. But back in 1956, two sociologists, Kingsley Davis and Judith Blake, argued that these big social forces never touch fertility directly. They always work through a smaller set of biological and behavioural channels, the “intermediate variables.” Their analytic framework remains one of the most cited starting points in modern demography, and it still helps explain why fertility falls in Kerala but stays higher in parts of Bihar.

What is the Davis and Blake model? 🔗

In their landmark 1956 paper Social Structure and Fertility: An Analytic Framework, Davis and Blake argued that every social, cultural, or economic factor influencing fertility must operate through one of eleven intermediate variables. They reasoned that for a live birth to occur, a woman must go through three sequential stages: sexual intercourse, conception, and a successful pregnancy. So, any factor affecting fertility has to act on at least one of these three stages.

This gave them a clean three-fold classification:

• Intercourse variables: factors that determine exposure to sexual intercourse.
• Conception variables: factors that determine whether intercourse leads to conception.
• Gestation variables: factors that determine whether a pregnancy ends in a live birth.

The brilliance of this framework lies in its logic. Education, income, caste, religion, or government policy cannot magically change a country’s birth rate. They first have to change something concrete, like the age at which women marry, the use of contraception, or the rate of induced abortion. Davis and Blake gave demographers a checklist of pathways to look at.

Background factors versus intermediate variables 🔗

The model distinguishes two types of factors that affect fertility: background or indirect variables, and intermediate or direct variables. Background factors include culture, religion, economy, and education. Intermediate variables are the biological and behavioural channels through which these background factors actually translate into births. If you change a background factor like female literacy, fertility will only fall if it shifts at least one intermediate variable, such as raising the age at marriage or increasing contraceptive use.

The eleven intermediate variables 🔗

Davis and Blake listed eleven variables across the three categories. Let’s break them down.

Intercourse variables (six factors) 🔗

These deal with how often, and for how long during the reproductive years, a woman is exposed to sexual intercourse. They split further into two groups: those that govern the formation and dissolution of sexual unions, and those that govern intercourse within a union.

Factors governing union formation and dissolution include:

• Age of entry into sexual unions: The earlier women enter marriage or cohabitation, the longer their reproductive exposure.
• Permanent celibacy: The proportion of women who never marry or enter a union throughout their reproductive years.
• Time spent after or between unions: Years lost to widowhood, divorce, or separation when women are not in a sexual union.

Factors governing exposure within unions include:

• Voluntary abstinence: Refraining from sex for religious, cultural, or personal reasons, including postpartum abstinence common in many traditional societies.
• Involuntary abstinence: Forced abstinence due to illness, impotence, long-term separation (such as labour migration), or imprisonment.
• Coital frequency: How often a couple has sex when they are together and able to.

Conception variables (three factors) 🔗

Even when intercourse takes place, conception is not guaranteed. Three intermediate variables decide this:

• Fecundity or infecundity from involuntary causes: Biological capacity to conceive, affected by nutrition, disease, breastfeeding-induced amenorrhea, and age.
• Use or non-use of contraception: All deliberate methods, mechanical, chemical, hormonal, or traditional, to prevent conception.
• Fecundity or infecundity from voluntary causes: Permanent steps like sterilisation or vasectomy that end the capacity to conceive.

Gestation variables (two factors) 🔗

Once conception happens, the pregnancy must end in a live birth for fertility to be realised:

• Foetal mortality from involuntary causes: Spontaneous miscarriage or stillbirth.
• Foetal mortality from voluntary causes: Induced abortion, whether legal or clandestine.

Together, these eleven variables form a comprehensive map of every biological and behavioural pathway through which fertility can rise or fall.

A closer look at the intercourse variables 🔗

The intercourse variables are arguably the most socially loaded of the three groups, because they are shaped almost entirely by cultural norms, laws, and economic conditions. They are also the most relevant for understanding fertility patterns across Indian states.

Age at marriage 🔗

This single variable explains a huge portion of fertility differences across communities. The legal age at marriage for women in India is 18, and a proposal to raise it to 21 has been under active discussion. The NFHS-5 report shows that the median age at first marriage for women aged 25 to 49 has been rising steadily over decades, accompanying a fall in fertility. A study using all five NFHS rounds found that around half the rise in the mean age at first cohabitation and first birth from 1992 to 2021 was due to changes in women’s characteristics like education and mass media exposure. Every additional year a young woman spends in school or work, rather than marriage, shortens her reproductive window and tends to reduce her completed family size.

Celibacy and remaining unmarried 🔗

Permanent celibacy is rare in most of South Asia, where marriage is near-universal. This is one reason why fertility in this region has historically been higher than in parts of Europe or Latin America, where larger shares of women never marry. However, recent decades have seen a rise in never-married women in urban metros, particularly among highly educated professionals, and demographers are beginning to track its effect on national fertility.

Marital dissolution and time between unions 🔗

Widowhood, divorce, and separation reduce the years a woman is exposed to intercourse within a union. In contexts where widow remarriage is socially discouraged, large numbers of women lose several reproductive years. Conversely, where remarriage is common and quick, this variable matters less. The taboo on widow remarriage in parts of India was historically a fertility-reducing factor among certain communities.

Voluntary and involuntary abstinence 🔗

Voluntary abstinence includes postpartum abstinence, religious fasts, and personal choice. Involuntary abstinence often arises from labour migration. When millions of men in states like Bihar, Uttar Pradesh, or Odisha migrate to other states for work, their wives back home effectively experience long periods of involuntary abstinence. This affects birth spacing in surprising ways and is increasingly important in studies of internal migration in India.

Coital frequency 🔗

This is the hardest variable to measure because it relies on self-reported data, but it matters. Couples living in joint families with little privacy, or those working in different cities, often have lower coital frequency, which lowers fertility independently of contraception.

How conception and gestation variables play out today 🔗

The conception variables are where modern family planning has had its biggest impact. NFHS-5 data show that the contraceptive prevalence rate among currently married women in India rose to about 67 percent in 2019 to 21, with female sterilisation remaining the dominant method. This single intermediate variable, the deliberate use of contraception, explains the bulk of India’s fertility decline from over 5 children per woman in the 1970s to a Total Fertility Rate below replacement today.

The fertility-inhibiting effect of infecundity is also visible: NFHS-5 data show a primary infertility rate of about 18.7 per 1,000 currently married women in union for at least five years, which means a small but meaningful share of couples never have a first birth despite trying. Gestation variables, particularly induced abortion under the Medical Termination of Pregnancy Act, also reduce the share of pregnancies that result in live births, though access remains uneven across states.

Why this model matters for fertility studies 🔗

The Davis and Blake framework gave demographers something they badly needed: a way to compare fertility across very different societies using a common analytical structure. Before this model, scholars often debated whether religion, caste, or income “caused” high fertility without specifying the mechanism. Davis and Blake forced everyone to ask a sharper question: through which intermediate variable does this background factor operate?

From background to behaviour: the policy lens 🔗

Policymakers find the model useful because background factors like literacy or per capita income cannot be changed overnight, but intermediate variables can be targeted directly. India’s family planning programme is essentially a policy intervention on one intermediate variable, contraception. Laws raising the minimum age at marriage target the age of entry into sexual unions. Female education works partly by delaying marriage and partly by increasing contraceptive use. The framework makes it possible to attribute fertility change to specific pathways.

The bridge to Bongaarts 🔗

In 1978, demographer John Bongaarts built directly on Davis and Blake’s work and reduced the eleven variables to four measurable proximate determinants: proportion married, contraceptive use and effectiveness, induced abortion, and postpartum infecundability (largely from breastfeeding). The Bongaarts model is now the workhorse of applied fertility analysis, but it would not exist without the conceptual map that Davis and Blake drew first. Anyone analysing NFHS data today is still, indirectly, using their classification.

Limitations of the model 🔗

The framework is not without critics. It treats lactation only as a part of involuntary infecundity, missing the powerful birth-spacing effect of prolonged breastfeeding that demographers now recognise as a major fertility regulator. It also focuses heavily on marriage and may underrepresent fertility within non-marital unions, which is increasingly relevant in urban contexts. Some sociologists have argued that the model is too institutional and pays less attention to individual decision-making or couple-level dynamics. Even so, it remains the foundation on which the entire field of proximate determinants is built.

Connecting back to socio-economic change 🔗

If you want to predict where fertility is heading in any state or community, the Davis and Blake checklist gives you a clean starting point. Look at age at marriage, look at contraceptive prevalence, look at migration patterns affecting spousal separation, look at access to abortion services. Each of these is a measurable channel through which deeper forces, female education, urbanisation, economic opportunity, and policy choices, leave their mark on the number of children born.

This is why fertility differences across Indian states map so closely onto these intermediate variables. Kerala’s low fertility is not just about literacy in the abstract; it is about late marriage, high contraceptive use, and small foetal loss. Bihar’s higher fertility reflects earlier marriage, lower contraceptive prevalence, and lower female autonomy. The background story is education and economy; the working machinery is the eleven variables.

What do you think? If India raises the legal age of marriage for women to 21, which other intermediate variables do you think will shift along with it, and which might resist change? And in a country where labour migration separates millions of couples for months at a time, how much of the recent fertility decline can really be attributed to family planning programmes versus involuntary abstinence?

References

1. https://www.nationalacademies.org/read/2207/chapter/4
2. https://grodri.github.io/demography/ProximateDeterminants.pdf
3. https://pmc.ncbi.nlm.nih.gov/articles/PMC7276374/
4. https://dhsprogram.com/pubs/pdf/FR375/FR375.pdf
5. https://pmc.ncbi.nlm.nih.gov/articles/PMC10061699/
6. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12487562/
7. https://pmc.ncbi.nlm.nih.gov/articles/PMC10757690/
8. http://papp.iussp.org/sessions/papp101_s05/PAPP101_s05_030_030.html
9. https://wfs.dhsprogram.com/WFS-CS/ISI-WFS_CS-39_Casterline%20et%20al_1984_The%20Proximate%20Determinants%20of%20Fertility.pdf