Regional Migration Flow Systems: Strategies for data assessment, validation and estimation

Diego Iturralde

Statistics SA

IFMS 2020

Layout

• Flow data vs stock data

• Data sources (Census, admin data, longitudinal data, big data)

• Collection of data and use in policy space

• Date of entry, country of origin

• Granularity of origin and destination

• Flow survey →World migration survey

• Case of South Africa

Flow data vs Stock data

• Migration data is often centred around the analysis of movement of people in the aggregate. The question being asked is often that of how many people have moved? From where and to where?

• Whilst this is a valid question the aggregate masks the issue of movement of people in a recent period? Have recent trends changed? Are there new patterns that we need to be made aware of?

• Most importantly what has the year on year change been that is able to inform policy, to update population estimates and to better inform

• Whilst stock data is widely available flow data is less commonly available

• Data sources for migration flow data can include but are not limited to the following

• Census questions which focus on movement since the last Census but are limited to the last move and are only collected every ten years.

• Administrative data in the form of human movement control type data at borders and entry points (in the case of international migration)

• Big data such as Facebook analytics shown by Spyratos et al

• In the case of administrative data, renewals such as residence permits are used as proxies for flows

• In many cases there is a focus on collecting inflow data but few countries count outflow data

Measurement of Migration in Census in SA

• Census is the most authoritative international migration data collection instrument which covers inflows to the country

• It measures migration with reference to a de facto household. Households comprise all resident members at the time of the census

• Focusses on last move only

• Undercount is adjusted for by means of a post-enumeration survey and amongst the major challenges leading to this undercount is access to high-walled areas

• P08(a) and P11 can be used in combination to measure flows although this cant be disaggregated by country of origin in the recent period but rather by country of birth

National Census 2011

National Census 2011

• Answering questions honestly for fear of victimisation

• Incorrect following of instruction with respect to usual resident

• Comments have been made about need for granularity of origin and destination. What are policy needs in this regard

• Migration is competing in space for questions in questionnaire

Flow Characteristics

• Based on 2011 Census broken down by those entering the country since 2006 and as a double check, those reporting they do not live in the same place as at the time of Census 2001

0

100000

200000

300000

400000

500000

600000

Recent Migration Flows by Province of DestinationTotal = 1024563

0

100000

200000

300000

400000

500000

600000

700000

Male Female

Recent Migration Flows by Sex Total = 1024563

Recent Migration Flows by Age

0 - 14 15 - 24 25 - 49 50 - 64 65 +

Total = 1024563

0

50000

100000

150000

200000

250000

300000

350000

400000

No schooling Some primary Completed primary Some secondary Grade 12/Std10 Higher

Recent Migration Flows by EducationTotal = 828550

Total = 894989

0

100000

200000

300000

400000

500000

600000

Employed Unemployed Discouraged work-seeker Other not economically active

Resent Migration Flow by Official Employment Status

Conclusions

• Almost all analysis and outputs are based on stock data and very little use is made of flows for international migration. For internal migration the case is the opposite

• Need to elevate the benefit of using flow data and how this creates a more relevant discourse in the policy space

• Potential to use various sources to estimate bilateral for which investment in emigration sources is required