Sample Registration System (SRS): Estimating Birth and Death Rates in India

How does a country of over 1.4 billion people figure out how many babies are born or how many people die each year? The answer lies in a quiet but powerful demographic surveillance machinery that has been running for more than five decades – the Sample Registration System (SRS). While the Census gives us a once-a-decade headcount, and the Civil Registration System logs individual births and deaths, it is the SRS that fills the gap in between, delivering reliable annual estimates of vital rates that shape everything from health budgets to welfare schemes.

Why India needed the SRS in the first place 🔗

In the years following independence, planners faced an awkward problem. The country had ambitious Five-Year Plans, but very little reliable data on how fast the population was growing or how many infants were dying. Registration of births and deaths was largely voluntary, patchy, and uneven across states. Without accurate vital statistics, policy was essentially flying blind.

To plug this gap, the Office of the Registrar General of India (RGI) launched a pilot scheme of sample registration in 1964-65 in a few selected states. The pilots worked. The system was rolled out on a full scale during 1969-70, and by the middle of 1971, the SRS had stabilised across the country. The Registration of Births and Deaths Act, 1969 ran in parallel, but because civil registration would take decades to mature, the SRS was treated as an interim measure. Six decades later, it remains very much alive and is now considered one of the largest demographic surveys in the world.

The shift from voluntary registration to sampling 🔗

The early planners considered several alternatives – a one-time retrospective survey, multi-round surveys, or strengthening civil registration directly. Each had drawbacks. A single survey suffered from recall errors. Multi-round surveys were expensive. Civil registration was structurally weak. The RGI settled on a clever middle path: take a representative sample of villages and urban blocks, and within those, use two independent observers to record every birth and death. This is the foundation of what is called the dual record system.

How the dual reporting system actually works 🔗

The SRS is built on a deceptively simple insight: if two people independently record the same events, you can compare their notes and arrive at a count that is closer to the truth than either alone. In practice, this involves four moving parts working in sequence.

1. Baseline survey of sample units 🔗

Before any continuous tracking begins, a baseline survey is conducted in each sample village or urban block. This survey identifies the usual resident population, prepares a houselist, and creates a list of women in the reproductive age group along with their pregnancy status. This forms the denominator for all rates that will follow.

2. Continuous enumeration by the part-time enumerator 🔗

The first observer is a part-time enumerator (PTE) – usually a primary school teacher, anganwadi worker, or similar local functionary who lives in the sample unit. Because the PTE is a resident, they are well-placed to hear about births and deaths as they happen. They record every vital event on an ongoing basis. As described in an overview of the SRS published in the Asian Population Studies literature, the enumerator covers rural households once every three months and urban households once a month, since the informant system is weaker in cities.

3. Independent half-yearly survey by a full-time supervisor 🔗

The second observer is a full-time supervisor, typically from the Directorate of Census Operations. Once every six months, this supervisor visits the same households and conducts an independent retrospective survey, asking about births and deaths that have taken place in the preceding six months. Critically, the PTE’s records are withdrawn from the field before this survey begins, so the supervisor’s count is completely independent.

4. Matching, verification, and reconciliation 🔗

This is where the magic happens. Events recorded by the PTE are matched event-by-event against the supervisor’s list. Three categories emerge – fully matched, partially matched, and unmatched. The partially matched and unmatched ones are then re-verified through a field visit, producing what statisticians call an unduplicated count. As the Madhya Pradesh census directorate notes in its description of SRS publications, this process not only catches errors of omission but also flags duplication, making the SRS a self-evaluating system.

What the SRS measures 🔗

From this carefully cleaned dataset, the Office of the Registrar General publishes an annual SRS Statistical Report and shorter SRS Bulletins. The key indicators include:

• Crude Birth Rate (CBR): live births per 1,000 population per year.
• Crude Death Rate (CDR): deaths per 1,000 population per year.
• Infant Mortality Rate (IMR): deaths of children under one year per 1,000 live births.
• Under-Five Mortality Rate (U5MR).
• Total Fertility Rate (TFR): average number of children per woman.
• Sex Ratio at Birth (SRB).
• Maternal Mortality Ratio (MMR), published in a special bulletin every few years.
• Life expectancy at birth, derived from abridged life tables.

These numbers are not just academic. The latest available SRS Statistical Report for 2023 showed CBR at 18.4, CDR at 6.4, and IMR at 25 per 1,000 live births. The report also noted Kerala continuing to have the lowest IMR at 5, while Madhya Pradesh, Chhattisgarh, and Uttar Pradesh shared the highest at 37 – a kind of state-level inequality that only SRS-style data can reveal year after year.

Why these numbers matter for policy 🔗

Programmes like the National Health Mission rely on SRS estimates to set targets, allocate funds, and measure progress. When the government claims that maternal mortality has dropped from 130 in 2014-16 to 93 in 2019-21, that figure comes from a Special Bulletin on Maternal Mortality based on SRS data. SRS numbers also feed into India’s reporting on the Sustainable Development Goals and into the Census-based population projections used by planners.

The strengths of the SRS 🔗

Five qualities have made the SRS the workhorse of Indian demography.

Reliable state-level estimates. Because the sample is large – currently covering more than 8,000 sample units spread across rural and urban India – the SRS can produce credible annual estimates not just for the country as a whole but for individual states, separately for rural and urban areas, by sex, and by broad age groups.

Annual frequency. Unlike the Census, which happens once every ten years, SRS data arrives every year. This makes it possible to track short-term trends and respond to demographic shifts in something close to real time.

Self-evaluating design. The dual reporting system contains its own quality check. The degree of mismatch between the PTE and the supervisor gives an internal measure of completeness, and reconciliation reduces both omissions and duplications.

Continuity over decades. The SRS has produced uninterrupted data since the early 1970s, making it one of the longest-running demographic surveillance systems anywhere in the world. This gives Indian researchers an unusually long time series to analyse.

Comparability with other sources. Because the methodology has remained broadly consistent, SRS estimates can be compared with results from the National Family Health Survey and the Census, helping triangulate the truth when sources disagree.

The weaknesses you should know about 🔗

For all its strengths, the SRS has real limitations that students of population studies must keep in mind.

It is a sample, not a count. The SRS does not cover every village or block; it covers a probability sample. That is enough for state and national estimates, but it cannot reliably produce district-level numbers. For granular sub-state planning, district administrators have to rely on the Civil Registration System or programme-based reporting like the Health Management Information System.

Dependence on incomplete civil registration. Although the CRS has improved dramatically and now captures over 98% of births and 97% of deaths, completeness varies sharply across states. In states with weak civil registration, the SRS is essentially the only credible source – which means an entire policy edifice rests on one survey system.

Time lag in publication. Annual reports historically take two to three years to appear. The 2023 report, for instance, was released in 2025. This lag, although shorter than in earlier decades, still limits the usefulness of SRS data for rapid response in fast-changing health situations.

Quality depends on the part-time enumerator. The whole continuous-enumeration arm rests on a local functionary working part-time. In areas where these workers are overburdened or where social conditions make enumeration difficult, both completeness and accuracy can suffer.

Internal inconsistencies and methodological debates. Researchers have flagged discrepancies between age-specific death rates and IMR figures in some recent SRS reports, raising questions about whether the system needs methodological refreshing. The Data for India explainer cited earlier notes an ongoing debate over which source – SRS or CRS – provides better estimates as civil registration matures.

Limited cause-of-death information. The verbal autopsy component, which assigns probable causes to deaths in the sample, is useful but cannot match the precision of medically certified causes of death.

The road ahead 🔗

The sampling frame of the SRS is revised once every ten years using the latest Census, with the most recent overhaul based on the 2011 Census. Future improvements will likely focus on tighter integration with the CRS, digital data capture, faster publication cycles, and richer socio-economic variables that allow researchers to study inequality in fertility and mortality more carefully. As civil registration approaches universal coverage in more states, the SRS may eventually evolve from being India’s primary source of vital statistics into a precision benchmark used to audit and validate administrative records.

The bigger picture 🔗

The SRS is a quiet success story of Indian statistical infrastructure. Born as a stopgap in the 1960s, it has outlived that label by half a century, providing the numbers that everyone from epidemiologists to economists to UN agencies routinely cite. Whenever you read a headline that says “India’s fertility rate has fallen below replacement level” or “infant mortality has dropped to a new low,” there is a very good chance the underlying figure was generated by an anganwadi worker noting a birth in a register, a supervisor cross-checking six months later, and a clerk in the Registrar General’s office matching the two records.

What do you think? If civil registration in India eventually reaches near-universal coverage, should the SRS be phased out as it was originally meant to be, or does an independent sample-based system still add something the CRS cannot? And given the time lag in SRS publication, how can India build faster vital-statistics pipelines for use in real-time health emergencies?

References

1. https://censusindia.gov.in/census.website/en/node/180
2. https://www.indiacode.nic.in/handle/123456789/1581
3. https://utkarsh.com/current-affairs/national/reports/indias-infant-mortality-rate-falls-to-25-srs-2023-report
4. https://www.researchgate.net/publication/326098440_The_Sample_Registration_System_SRS_in_India_An_Overview_as_of_2017
5. https://censusmp.gov.in/censusmp/english/srs.html
6. https://www.drishtiias.com/daily-updates/daily-news-analysis/sample-registration-system-srs-statistical-report-2023
7. https://nhm.gov.in/index1.php?lang=1&level=1&sublinkid=794&lid=168
8. https://www.pib.gov.in/PressReleaseIframePage.aspx?PRID=2128024&reg=3&lang=2
9. https://www.dataforindia.com/crs-srs-explainer/