Japanese inter-regional migration patterns affected by 2011 Tohoku Disaster, analyzed with 2015 Japan Population Census Prof. Makoto, Okumura International Research Institute of Disaster Science IRIDeS, Tohoku University, Sendai, Japan E-mail:mokmr@m.tohoku.ac.jp 1
Swift restoration gathers peoples attention, adding to the direct impact of the disaster Activity Level of Socio-economic Hazard and Exposure Vulnerability Loss Resilience Time Impact HazardExposureVulnerability We conduct a research to capture the restoration trend through population change based on the Annual migration statistics (Quantitative Analysis) Presented at IDRiM 2017 in Reykjavik (Aug.25) 2
Migration flows to be enlarged after a disaster! Hyogo Prefecture with Kobe Earthquake in 1995 350000" 300000" 250000" 200000" 150000" 100000" 50000" 0" Emigration Peaks in 94 95FY Birth_Hyg" Immig_Hyg" 1996FY Immigration Peak Hyogo%Pref. Death_Hyg" Emig_Hyg" 1979" 1981" 1983" 1985" 1987" 1989" 1991" 1993" 1995" 1997" 1999" 2001" 2003" 2005" 2007" 2009" 2011" 2013" 3
Statistical Analysis ofmigration flows 4 Annual Emigration / Immigration Numbers from/to each Prefecture Inhabitant Registration Statistics1973-2013 Migration free from disaster Normal Migration Pattern Enlargements by Disaster(s) regressed Fixed Effect of Panel Data Analysis Disaster Size (measured by annual R.I.A.: Relative Affected Inhabitants) Disaster Statistics by Fire Agency
Categorization of Disaster Size ( ) D i,t 5 Disaster size is captured by annual relative affected inhabitants. (Number of affected inhabitants / prefecture inhabitants) Variables are categorized in four different size. Annual R.A.I Number of cases Cumulative % No Disaster 0 296 14.35% Small Size D. 0 857 55.89% Medium Size D. 0.0001 587 84.34% Large Size D. 0.001 276 97.77% Huge Size D. 0.01 47 100.00% R.A.I is prepared annually for each prefecture (1970-2013) excluding Okinawa before 1974. We estimate the effect up to 3 years before. We did not consider Tohoku 3 Prefectures, because of lack of reliable number of affected people in Fire Agency Statistics.
1.4 1.3 1.2 1.1 1.0 0.9 0.8 0.7 0.6 Small same yr Small 1 yr before Estimated Enlargement effects by disaster of different size Small 2 yr before Emig Small 3 yr before Medium same yr Medium 1 yr before Immig R.I.A<0.01 Medium 2 yr before Medium 3 yr before Large same yr Large 1 yr before Emigration will shrink after one year No negative effect on Immigration No social decrease of Population Large 2 yr before Large 3 yr before R.I.A>0.01 Huge same yr Huge 1 yr before Huge 2 yr before Huge 3 yr before Emigration increase and Immigration decrease Social decrease of population 6 Effects of 2011 Disaster were not included!
Migration in Fukushima seems different after the Tohoku Earthquake, Tsunami and Nuclear Accident in 2011 160000" 140000" 120000" 100000" 80000" 60000" 40000" 20000" 0" Miyagi&Pref. Birth_Myg" Immig_Myg" 1979" 1981" 1983" 1985" 1987" 1989" 1991" 1993" 1995" 1997" 1999" 2001" 2003" 2005" 2007" 2009" 2011" 2013" Immigration Peak appeared, as well as Emigration Death_Myg" Emig_Myg" 120000" 100000" 80000" 60000" 40000" 20000" 0" Fukushima)Pref. Birth_Fksm" Immig_Fksm" Death_Fksm" Emig_Fksm" 1979" 1981" 1983" 1985" 1987" 1989" 1991" 1993" 1995" 1997" 1999" 2001" 2003" 2005" 2007" 2009" 2011" 2013" Only emigration Peak was appeared in Fukushima Let us investigate migration pattern more qualitatively! 7
TWO NATIONAL STATISTICS Migration Report by Residence Registration Japan Population Census Interval Annual 5 years (Questions on migration: once in 10 years) Published Contents Source Problem in Fukushima case Quantitative Info: Numbers of total emigrants and Immigrants of each Prefecture or municipality. Aggregate the residence registration from municipal governments Many emigrants do not move registration, in order to keep rights to get support for the people affected by the Nuclear Accidents. Including Qualitative Info.: Age Structure Origin and Destination Prefectures Exhaustive Survey to the people actually reside. (Direct distribution and collection) We cannot capture people s migrations who already return to Fukushima by Oct.2015. 8
NET-MIGRATION ANALYSIS For Age-classes 10000 5000 Fukushima Prefecture 5 years net-migration (Imigration emigration) Positive (gain) 0-5000 5~ 10~ 15~ 20~ 25~ 30~ 35~ 40~ 45~ 50~ 55~ 60~ 65~ 70~ 75~ 80~ 85~ -10000-15000 -20000 Negative (loss) 5 year age categoryage in period end year 19952000 20052010 20102015 2
Emigration from each Prefecture in the four intervals 10 1600000 1400000 1200000 1000000 800000 1985-90 1985-1990 1995 00 2005 10 1995-2000 2005-2010 2010-2015 600000 400000 200000 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Fukushima is the only prefecture where number of emigrants grew in between 2010-2015.
Immigration to each Prefecture in the four intervals 11 1,600,000 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 1975-80 1985-90 1995 00 2005 10 2010 15 1975-1980 1985-1990 1995-2000 2005-2010 2010-2015 200,000 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 There are no particular characteristics in Tohoku Prefectures
Net Immigration to each Prefecture in the four intervals 12 200000 150000 100000 1985-1990 1995-2000 2005-2010 2010-2015 1985-90 1995 00 2005 10 2010 15 50000 0-50000 -100000 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47-150000 -200000 Fukushima is the only prefecture where net loss of migration grew in between 2010-2015.
Qualitative Analysis focusing on gender and age. NMF: Non-negative Matrix Factorization 13 Non negative factorization Method can show several sex-age migration patterns and give weightings for each region. Each factor may correspond to objective of migration, but we do now know the number of factors a priori. n:regions x 2 msex & Age XE XI = k:factors n:regions x 2 WE WI k:factors msex & Age F How different age-groups jointly move with different purpose? Next, I show the case of 7 factors division.
9.00E-02 Factors: sex and age structure MiddleWorker1 DependentFamily2 MiddleFamily3 Collage4 SingleWorker5 NewJobWorker6 SingleWorker7 14 8.00E-02 7.00E-02 6.00E-02 5.00E-02 4.00E-02 3.00E-02 2.00E-02 1.00E-02 0.00E+00 M00 M05 M10 M15 M20 M25 M30 M35 M40 M45 M50 M55 M60 M65 M70 M75 M80 Male M85+ F00 F05 F10 F15 F20 F25 F30 F35 F40 F45 F50 F55 F60 F65 F70 F75 F80 F85+ Female
Scores: What s happen in eachprefecture? 15 SingleWorker7 NewJobWorker6 SingleWorker5 Collage4 MiddleFamily3 DependentFamily2 MiddleWorker1 2500000 2000000 1500000 Emigration Immigration 1000000 500000 0 Emig01 Emig03 Emig05 Emig07 Emig09 Emig11 Emig13 Emig15 Emig17 Emig19 Emig21 Emig23 Emig25 Emig27 Emig29 Emig31 Emig33 Emig35 Emig37 Emig39 Emig41 Emig43 Emig45 Emig47 Immig02 Immig04 Immig06 Immig08 Immig10 Immig12 Immig14 Immig16 Immig18 Immig20 Immig22 Immig24 Immig26 Immig28 Immig30 Immig32 Immig34 Immig36 Immig38 Immig40 Immig42 Immig44 Immig46
Weightings: What s happen in each prefecture? 16 SingleWorker7 NewJobWorker6 SingleWorker5 Collage4 MiddleFamily3 DependentFamily2 MiddleWorker1 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Emig01 Emig03 Emig05 Emig07 Emig09 Emig11 Emig13 Emig15 Emig17 Emig19 Emig21 Emig23 Emig25 Emig27 Emig29 Emig31 Emig33 Emig35 Emig37 Emig39 Emig41 Emig43 Emig45 Emig47 Immig02 Immig04 Immig06 Immig08 Immig10 Immig12 Immig14 Immig16 Immig18 Immig20 Immig22 Immig24 Immig26 Immig28 Immig30 Immig32 Immig34 Immig36 Immig38 Immig40 Immig42 Immig44 Immig46
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% Weightings (relative) in Northeastern 10 prefectures Emigration Immigration SingleWorker7 NewJobWorker6 SingleWorker7 NewJobWorker6 SingleWorker5 Collage4 SingleWorker5 Collage4 MiddleFamily3 DependentFamily2 MiddleFamily3 DependentFamily2 MiddleWorker1 MiddleWorker1 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 17 0% 01Hokkaido 02Aomori 03Iwate 04Miyagi Dependent Family 2 moved out from Fukushima 05Akita 06Yamagata 07Fukushima 08Ibaragi 09Tochigi 10Gumma 0% 01Hokkaido 02Aomori 03Iwate 04Miyagi 05Akita Middle Worker 1, New Job Worker 6 moved in to Fukushima 06Yamagata 07Fukushima 08Ibaragi 09Tochigi 10Gumma
9.00E-02 8.00E-02 7.00E-02 Strongly observed Factors in Fukushima MiddleWorker1 DependentFamily2 MiddleFamily3 Collage4 SingleWorker5 NewJobWorker6 SingleWorker7 Young Workers (more male than 18 6.00E-02 5.00E-02 4.00E-02 3.00E-02 New Collage Students went out Dependent Family as mothers and children went out 2.00E-02 1.00E-02 0.00E+00 M00 M05 M10 M15 M20 M25 M30 M35 M40 M45 M50 M55 M60 M65 M70 M75 M80 Male M85+ F00 F05 F10 F15 F20 F25 F30 F35 F40 F45 F50 F55 F60 F65 F70 F75 F80 F85+ Female
Fleshly gained results 19 We adopt the NMF analysis on a dataset of age-sex matrix prefecture versus immigrants, emigrants and stay in three periods. 1995-2000 (including Kobe earthquake?) 2005-2010 2010-2015 (including effect of 2011)
2010-15 Immigration Factor Scores Young Family Immigrants are observed anywhere and any time,. Only Exceptional Case is Fukushima 2010-2015. 20
21 Factor scores in Fukushima in 3 periods Emigration Immigration lincrease in Young Family ldisappear of Young Family s immigration. lnew Job Immigration increased 1995-00 2005-10 2010-15 1995-00 2005-10 2010-15
KOBE CASE HYOGO Pref. 22 100% Emigration Emigration Immigration Stay (not move) Immigration 80% 60% 40% 20% 0% 2000 2010 2015 2000 2010 2015 2000 2010 2015 New Job Workers, Mobility was relatively high in 1995-2000.
Discussion on the result 23 Unique patterns were observed in Fukushima. Young Mothers and children decided to emigrate from Fukushima (Dependent Family 2) Fear of influence of radioactivity. Middle aged (mostly male) workers (Middle Worker1) and young single workers finding the job (New Job Workers6) moved in Fukushima In order to have works such as disaster recovery or the radioactive decontamination. Middle Family and Dependent Family did not enter Fukushima Many households were dismantled.
24
FINDINGS 25 We investigated the effect of disaster on the inter-regional migration in the following 3 years, considering the disaster size. Small, Meduim, Large Disaster R.I.A.<0.0198% cases Emigration will shrink after one year No negative effect on Immigration No social decrease of Population External Assistance is not always necessary Huge Disaster R.I.A.>0.012% of the total cases Emigration increase and Immigration decrease Social decrease of population External Assistance is strongly necessary
Future Research Issue Consider Disaster type categorization Consider the external monetary assistance or designation of Serious Disaster to be supported Closer investigations, based on the smaller local area data such as municipalities, or different age groups Analysis of the effect on economic performance indexes 26