UNFPA/UNECE/NIDI Training programme on international migration, Geneva, 24-28/01/2005 Analytical Measures and Methods George Groenewold, NiDi 1. Demographic Balance Equation: 2.Method related concepts and definitions 3.Main approaches: A. Migration status is known: direct methods A.1Cross-tabulation migration status characteristics and an example B. Migration status is unknown: indirect methods B.1Vital Statistics or Residual Method and an example B.2Intercensal Cohort Component Method and an example

1. The Demographic Balance Equation P(t+1)= P(t)+Births-Deaths+(Immigrants-Emigrants), or, P(t+1)= P(t)+Natural Increase+Net Migration

Out-migration (emigration) rate: number of persons that moved out of place X divided by the average number of persons that were exposed to migration In-migration (immigration) rate: number of persons that live in place x, but originated from elsewhere, divided by the average number of persons that were exposed to migration Net migration: the difference between the number of immigrants (or in- migrants) and the number of emigrants (or out-migrants) Note:no concensus about which persons should be in the denominator in the above measures (UN, 1970, pp40-42.), depends on research question which one to choose Lifetime migrant: someone who at the time of the census lives away from his/her place of birth 2. Method related concepts and definitions

Cohort A cohort is a group of individuals experiencing the same demographic (vital) event. Birth cohort=persons born in the same year. Migration cohort=persons who emigrated from country X in the same year. Life table. A life-table summarizes the (age specific) mortality experience in a population in a particular year or particular (intercensal) period. Separate life tables are constructed for men and women. A full life-table can be derived from a set of Age Specific Mortality Rates (ASMR’s). Two important products of the life table are: (1) the life expectancy at birth and (2) age-specific survival ratio’s Survival Ratio.Indicator for the probability of surviving of persons in a particular age-group X to the next higher age-group X+5. Survival ratio’s are used in cohort component projections of the population Cohort Component Projection Method. Estimation method that builds on the principle that the size and age-structure of the current population can be predicted from an age distribution in the past if interperiod births, deaths and net numbers of (international) migrants are known

3. Main approaches: A. Migration status is known: direct method (D) A.1 Cross-tabulation of migration status characteristics Place of Usual Residence (POR) by Place of Birth -life time migration moves only, not time-specific -severe underestimation migration moves (intermediate destinations not recorded) POR by POR 1 or 5 years agoExample -focus on recent migration flows -UN recommends 5 years ago -recall problem with fixed reference date POR by Place of Previous Residence by Duration of Residence -refined migration cohort analysis, and migration and development analysis - recall problem

Table 1.Place of (usual) residence at time of census by Place of residence 1 year ago

Table 1. (continued)

B. Migration status is unknown: indirect methods B.1Vital Statistics or Residual Method and an example Demographic Balance Equation P(t+1)= P(t)+Births-Deaths+(Immigrants-Emigrants), or, P(t+1)= P(t)+Natural Increase+Net Migration or, Net Migration=P(t+1) - (P(t)+Natural Increase) Indirect estimation: Migration estimate derived as ‘residual’ from two subsequent census populations and intercensal natural increase (=recorded or estimated intercensal births-deaths)

Table 2.Derivation of net migration estimates if population figures at two points in time are known and numbers of births and deaths are known between these points in time

Table 2. (continued)

B.2 Intercensal Cohort Component Method and an example Data needs - Age distributions by sex of two consecutive censuses - intercensal life-tables by sex - Estimate of intercensal fertility (ASFR by sex, or ASFR and sex ratio at birth The principle 1.Project age groups (0-75 years) in first census, in year t, to the date of the second census using appropriate intercensal life tables for each 5- year projection interval: P(t,x) × S(x, x+5)=P(t+5,x+5), where x=age-group 2.Estimate births in 5-year projction interval and project number of surviving births in each 5-year projection interval 3.Repeat (1. ) and (2.) for each five-year projection interval and substract the projected population at the time of the second census from the observed census population to obtain age specific and total net migration estimate (=indirect or residual estimation)

Figure 1.Lexis-diagram, illustrating the principle of Cohort Component Projections

Step 1: derivation/selection of an appropriate life table

Step 2: Estimation of Births and births surviving projection interval

Step 3:Derivation of age-specific net-migration estimates, assuming that second census is five-years later than first census

If census is 10 years apart, add another projection interval, I.e. repeat steps 2 and 3 to determine projected population 10 years later. Use different fertility and/or mortality rates if appropriate This approach may slightly over- or underestimate the above age- specific net migration numbers Solution: 1.Reverse survive the survivors observed in the second census to the date of the first census and derive estimate of net number of migrants from that projection. 2.Average the forward and reverse-survived estimate of age-specific net migration For Cohort Component Projections, use software, e.g.: - PEOPLE-FiVFiV - PAS-LIPRO - DemProj- PDPM_PC- MortPak-Lite

Suggested literature.