Migration Estimation: Understanding Patterns, Methods, and Impact

Migration is one of the three pillars of population change, sitting alongside fertility and mortality. But unlike births and deaths, which are discrete and largely one-time events, migration is messy. A person can move multiple times in a lifetime, across short or long distances, for weeks or for decades, and often without leaving any official record. This makes migration estimation one of the most fascinating yet challenging tasks in population studies. To understand how demographers tackle this puzzle, we need to look at what migration really means, what dimensions shape its measurement, what drives people to move, and how researchers actually count those who often go uncounted.

Migration as a complex phenomenon 🔗

At its simplest, migration is the movement of people from one place of residence to another. The International Organization for Migration defines it as movement away from a usual place of residence to a new one, either across an international border or within a country. But this clean definition hides layers of complexity.

Migration is not just a geographic act. It is simultaneously a socio-economic phenomenon, where households weigh wages, costs, and opportunities; a political phenomenon, shaped by border policies, citizenship rules, and conflict; and a cultural phenomenon, influenced by language, kinship networks, and the social acceptance of leaving one’s home. A construction worker moving from Bihar to Delhi for six months of work, a bride relocating to her husband’s village after marriage, a student moving to Bengaluru for a degree, and a software engineer migrating to Canada on a permanent residency visa are all classified as migrants, yet their experiences and the data needed to capture them differ enormously.

This is precisely why scholars argue that no single definition of migration suits every purpose. Each researcher has to specify what counts as “significant” distance and “permanent” duration based on the question they want to answer.

Components of migration measurement 🔗

To convert this complex phenomenon into countable numbers, demographers break migration down into measurable components. Four dimensions stand out: duration, distance, origin, and destination.

Duration of the move 🔗

How long must a person stay in the new place before being called a migrant? The answer is not obvious. A visit lasting a weekend is not migration; a permanent shift clearly is. Everything in between requires a rule. The United Nations Department of Economic and Social Affairs generally distinguishes between short-term or temporary migration, covering movements of three to twelve months, and long-term or permanent migration, defined as a change of country of residence for at least one year. Under the 2021 revision of international migration statistics, a person is considered a resident migrant if they have been living in a country for most of the last twelve months, while shorter movers between three and twelve months are now categorised as “movers” rather than migrants.

Duration matters because it determines who appears in the data. Censuses that ask only about long-term residence systematically miss seasonal labourers, who form a huge share of internal movement in countries like India.

Distance of the move 🔗

Distance is the second yardstick. Shifting houses within the same neighbourhood is residential mobility, not migration. Moving to a different district, state, or country is. In India, the Census classifies internal movement into intra-district, inter-district, and inter-state streams, with intra-district migration making up the largest share at around 21.8% of the population in 2011, followed by inter-district at 10% and inter-state at 4.7%.

The further a person moves, the bigger the social, economic, and cultural shift, and the more likely it is to appear in formal records. But short-distance movement, though less visible, is often more frequent and more relevant for local planning.

Origin and destination 🔗

Every migration event has a “from” and a “to”. The origin is the place of last residence or the place of birth; the destination is where the person is enumerated. The Indian Census collects migration data using two criteria, place of birth and place of last residence, and the latter is generally preferred because it captures more recent moves. Combining the two allows demographers to construct migration streams, the four classic flows being rural-to-rural, rural-to-urban, urban-to-rural, and urban-to-urban.

According to the 2011 Census, migrants based on place of last residence totalled 453 million, accounting for 37.5% of the total population, twice the number recorded in 1971. Rural-to-urban migration accounted for nearly 48.5% of the stream in 2011, driven largely by employment opportunities in expanding cities.

Determinants of migration 🔗

Why do people move? Migration estimation is not just about counting bodies; it is also about understanding the characteristics that shape who migrates, when, and why. Several determinants feed directly into how migration is measured and interpreted.

Age 🔗

Migration is highly age-selective. Young adults migrate far more than children or the elderly because they are at the stage of life when education, work, and marriage create reasons to move. A study of Indian migration patterns using the 2011 Census found that Indians display early migration profiles, with peaks at around age 20 for females and age 25 for males. These age peaks are so consistent across populations that demographers have built model migration schedules to describe them, allowing them to fill in gaps in poor-quality data.

Sex 🔗

Men and women migrate for very different reasons, and this shapes how their movements are measured. In India, female migration is dominated by marriage, while male migration is driven by employment. The 2011 Census showed female migration rates rising from 41.64 to 52.17 per 100 between 1991 and 2011, while male rates rose from 14.79 to 23.03. Without disaggregating by sex, an analyst would misread both the volume and the meaning of migration in any region.

Socio-economic status 🔗

Income, education, caste, and occupation strongly influence who migrates and where. Poorer states like Bihar, Uttar Pradesh, Jharkhand, and Odisha, which belong to the Empowered Action Group, are major sources of out-migration, while richer, more urbanised states like Maharashtra, Delhi, and Gujarat are major destinations. Migration in this sense becomes a labour market response to inter-state inequality, and any estimate that ignores socio-economic conditions cannot explain the patterns it observes.

Purpose of migration 🔗

Censuses and surveys also capture why people move. Common categories include work or employment, business, education, marriage, movement with the household, and “others” such as natural disasters or political displacement. The mix of reasons is a vital input for measurement because different purposes correspond to different durations, frequencies, and distances. Migration data experts further distinguish labour migration, forced displacement, irregular migration, and environmental migration, each of which requires tailored measurement strategies.

Direct and indirect methods of migration estimation 🔗

Once the dimensions and determinants are clear, the question becomes practical: how do demographers actually estimate migration? Broadly, methods fall into two families, direct and indirect.

Direct methods 🔗

Direct methods rely on data that explicitly records movement. The three main sources are population censuses, sample surveys, and continuous population registers. Vital and population registers, where every change of residence is recorded as it happens, are considered the most accurate source where they exist, mainly in developed countries with strong registration systems.

Where registers are weak, censuses do most of the work. They ask questions like place of birth, place of last residence, duration of stay, and reason for migration. In India, both the Census and the National Sample Survey Office collect such data. Strengths of direct methods include actual recorded events, the ability to disaggregate by age, sex, and reason, and full geographic coverage in a census. Limitations include the fact that censuses occur only once a decade, that they tend to miss short-term, seasonal, and circular migration, and that respondents may not recall move details accurately. Researchers note that surveys often fail to record temporary, seasonal, and circulatory movements, which are central features of Indian labour mobility.

Indirect methods 🔗

When direct data is missing, incomplete, or outdated, demographers turn to indirect methods. These techniques infer migration from other demographic information, drawing on what we already know about births, deaths, and population growth. The United Nations 1983 manual on indirect estimation described these techniques as “suited for analysis of incomplete or defective demographic data”, capturing both their purpose and their limits.

The simplest and most widely used indirect approach is the vital statistics method based on the demographic balancing equation. The idea is intuitive: any change in a population between two points in time must come from either natural increase (births minus deaths) or net migration (in-migrants minus out-migrants). So if we know the population at two dates and the births and deaths in between, the leftover difference must be net migration. Indirect methods of this kind rest on the principle that a population changes through births, deaths, and migration, and net migration can be recovered as the residual.

A second family of techniques uses survival ratios. The census survival ratio method, for instance, takes a cohort enumerated in one census, applies an expected survival ratio to predict how many should remain in the next census, and treats the difference between expected and observed numbers as net migration for that cohort. Variants like the forward and reverse census survival ratio methods allow age-specific estimates of net migration, which are particularly useful where vital registration is weak.

More advanced indirect techniques use model schedules and log-linear models to estimate age patterns and spatial flows of migration where data is irregular or partial. Even questions to mothers about the residence of their children, or to respondents about the residence of their siblings, have been used in some surveys to estimate emigration in countries where direct records are missing. The PubMed-indexed review of indirect estimation of migration notes that such techniques were originally developed by an IUSSP working group to complement traditional data on place of birth and duration of residence.

Indirect methods have the obvious advantage of producing estimates where none would otherwise exist. Their main weakness is that they typically yield only net migration, not gross flows; they cannot tell us how many people moved in or out separately, nor why they moved.

Why migration estimation matters 🔗

Migration estimates are not academic curiosities. They drive decisions about how many schools and hospitals a city will need, where to invest in housing and transport, how to design social protection for circular workers, and how to respond to crises such as the reverse migration witnessed during the COVID-19 lockdown. Underestimating migration leaves cities under-resourced and migrants invisible; overestimating it distorts planning in source regions. Getting both the concept and the method right is therefore central to making population science useful for policy.

What do you think? Given that seasonal and circular migrants in India often slip through the gaps of both censuses and surveys, what kinds of new data sources, ranging from mobile phone records to migrant welfare registrations, might give us a fuller picture of who is moving and why? And when direct and indirect methods give conflicting estimates for the same region, which one would you trust more, and on what grounds?

References

1. https://www.iom.int/fundamentals-migration
2. https://www.encyclopedia.com/social-sciences-and-law/sociology-and-social-reform/sociology-general-terms-and-concepts/migration
3. https://refugeesmigrants.un.org/definitions
4. https://www.migrationdataportal.org/handbooks/chapter-1-key-concepts-migration-data-and-their-measurement/international-mobility
5. https://iasp.ac.in/uploads/journal/10.%20Migration%20in%20India%20trends%20and%20characteristics-1669206793.pdf
6. https://www.migrationletters.com/index.php/ml/article/view/10538
7. https://www.researchgate.net/publication/343093233_Internal_Migration_in_India
8. https://mpra.ub.uni-muenchen.de/61577/1/MPRA_paper_61577.pdf
9. https://www.migrationdataportal.org/handbooks/chapter-4-introduction-key-concepts-and-definitions
10. https://www.internationalscholarsjournals.com/articles/direct-and-indirect-methods-of-demography.pdf
11. https://www.worldwidejournals.com/indian-journal-of-applied-research-(IJAR)/recent_issues_pdf/2020/May/an-overview-of-trends-and-patterns-of-internal-migration-in-india-2011_May_2020_1588325627_9408784.pdf
12. https://link.springer.com/book/10.1007/978-90-481-8915-1
13. https://link.springer.com/rwe/10.1007/978-94-007-6179-7_83-1
14. https://pubmed.ncbi.nlm.nih.gov/12280919/