Taking the Skill Bias out of Global Migration Costanza Biavaschi - PowerPoint PPT Presentation

Calibration 1) & 2) Values from previous lit + match moments Parameter Value Source Preference parameters 0.5 exogenous β β T 0.135 calibrated (match consumption to production) 3 exogenous θ 0.5 exogenous µ 4 Simonovska (2014) ε 5 Docquier, Özden & Peri (2013) σ s 20 Ottaviano & Peri (2013) σ n Worker efficiency parameters a F 0.478 calibrated to match OECD average i a L 0.12-0.40 calibrated from FOC of cost minimization i a H 0.24-0.60 calibrated from FOC of cost minimization i

Calibration 1) & 2) Values from previous lit + match moments Parameter Value Source Preference parameters 0.5 exogenous β β T 0.135 calibrated (match consumption to production) 3 exogenous θ 0.5 exogenous µ 4 Simonovska (2014) ε 5 Docquier, Özden & Peri (2013) σ s 20 Ottaviano & Peri (2013) σ n Worker efficiency parameters a F 0.478 calibrated to match OECD average i a L 0.12-0.40 calibrated from FOC of cost minimization i a H 0.24-0.60 calibrated from FOC of cost minimization i 3) Find GDP pc and trade costs : iterative procedure Appendix: extensive sensitivity checks

Welfare gains/losses from skill-bias = U baseline − U counterfactual ∆ U U U counterfactual

Change in welfare in % -15 -10 -5 0 Haiti Welfare effects - non-OECD countries Jamaica Albania Morocco El Salvador Zimbabwe GHANA GHANA GHANA GHANA GHANA GHANA Senegal Uruguay Moldova SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA Tunisia Philippines MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE Zambia Colombia Romania Vietnam INDIA INDIA INDIA INDIA INDIA INDIA Bangladesh BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL CHINA CHINA CHINA CHINA CHINA CHINA Mexico Ukraine Chile Bulgaria

Welfare, whose welfare? Problem: base population changes!

Welfare, whose welfare? Problem: base population changes! Effect a mixture of "treatment" and "composition" effect ◮ Treatment effect: on non-migrants ◮ Composition effect: on migrants

Welfare, whose welfare? Problem: base population changes! Effect a mixture of "treatment" and "composition" effect ◮ Treatment effect: on non-migrants ◮ Composition effect: on migrants Solution: welfare per never-migrant

Change in welfare in % -15 -10 -5 Welfare effects - selected sending countries 0 Haiti Jamaica Albania Morocco El Salvador Zimbabwe GHANA GHANA GHANA GHANA GHANA GHANA Senegal Uruguay Moldova SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA Tunisia Philippines MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE Zambia Colombia Romania Vietnam INDIA INDIA INDIA INDIA INDIA INDIA Bangladesh BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL CHINA CHINA CHINA CHINA CHINA CHINA Mexico Ukraine Chile Bulgaria

Change in welfare in % -15 -10 -5 Welfare effects - selected sending countries 0 Haiti Jamaica Albania Morocco El Salvador Zimbabwe GHANA GHANA GHANA GHANA GHANA GHANA Senegal Uruguay Moldova Welfare per capita Welfare per never-migrant SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA Tunisia Philippines MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE Zambia Colombia Romania Vietnam INDIA INDIA INDIA INDIA INDIA INDIA Bangladesh BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL CHINA CHINA CHINA CHINA CHINA CHINA Mexico Ukraine Chile Bulgaria

Change in welfare in % -2 0 2 4 6 Iceland Germany Estonia Welfare effects - OECD countries Italy FINLAND FINLAND FINLAND Austria Belgium Slovenia Denmark Spain Welfare per capita Welfare per never-migrant Greece France Portugal Norway Sweden Netherlands New Zealand Ireland United Kingdom Switzerland UNITED STATES UNITED STATES UNITED STATES Australia Luxembourg Israel CANADA CANADA CANADA

Global effects 3 Change in welfare in % 2 1 0 -1 WORLD OECD NON-OECD Welfare per capita Welfare per never-migrant

How important in sending countries?

Change in welfare in % -20 -15 -10 -5 0 Haiti How important in sending countries? Jamaica Albania Morocco El Salvador Zimbabwe Welfare effect - current migration vs zero migration Welfare effect of skill bias GHANA GHANA GHANA GHANA GHANA GHANA Senegal Uruguay Moldova SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA Tunisia Philippines MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE Zambia Colombia Romania Vietnam INDIA INDIA INDIA INDIA INDIA INDIA Bangladesh BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL CHINA CHINA CHINA CHINA CHINA CHINA Mexico Ukraine Chile Bulgaria

Change in welfare in % 10 15 20 0 5 Iceland How important in receiving countries? Germany Estonia Italy FINLAND FINLAND FINLAND Austria Welfare effect - current migration vs zero migration Welfare effect of skill bias Belgium Slovenia Denmark Spain Greece France Portugal Norway Sweden Netherlands New Zealand Ireland United Kingdom Switzerland UNITED STATES UNITED STATES UNITED STATES Australia Luxembourg Israel CANADA CANADA CANADA

How important globally? 6 Change in welfare in % 4 2 0 -2 WORLD OECD NON-OECD Welfare effect - current migration vs zero migration Welfare effect of skill bias

Distributional effects Global 3 Change in real wages in % 2 1 0 -1 WORLD OECD NON-OECD Real wages low-skilled non-migrants Real wages medium-skilled non-migrants Real wages high-skilled non-migrants

Further extensions Why positive for low-skilled? Remittances Brain gain effect TFP externality (Lucas, 1988) Migrant networks and trade Downskilling

Most plausible scenario Externality Parameter Minimalist Intermediate Maximalist Remittances γ 0 0.5 1 Brain gain σ b 0.01 0.02 0.05 TFP σ a 0.1 0.3 0.5 Network effects σ t 0 -0.02 -0.04 Downskilling - yes no no

Sending countries Change in welfare in % -6 -4 -2 0 2 Haiti Jamaica Albania Morocco El Salvador Most plausible scenario Zimbabwe GHANA GHANA GHANA GHANA GHANA GHANA Welfare effect with all extensions - maximalist Welfare effect with all extensions - intermediate Welfare effect with all extensions - minimalist Welfare effect - baseline Senegal Uruguay Moldova SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA Tunisia Philippines MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE Zambia Colombia Romania Vietnam INDIA INDIA INDIA INDIA INDIA INDIA Bangladesh BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL CHINA CHINA CHINA CHINA CHINA CHINA Mexico Ukraine Chile Bulgaria

Receiving countries Change in welfare in % -1 0 1 2 3 4 Iceland Germany Estonia Italy FINLAND FINLAND FINLAND Most plausible scenario Austria Welfare effect with all extensions - maximalist Welfare effect with all extensions - intermediate Welfare effect with all extensions - minimalist Welfare effect - baseline Belgium Slovenia Denmark Spain Greece France Portugal Norway Sweden Netherlands New Zealand Ireland United Kingdom Switzerland UNITED STATES UNITED STATES UNITED STATES Australia Luxembourg Israel CANADA CANADA CANADA

Most plausible scenario Global 1 Change in welfare in % .5 0 -.5 WORLD OECD NON-OECD Welfare effect - baseline Welfare effect with all extensions - minimalist Welfare effect with all extensions - intermediate Welfare effect with all extensions - maximalist

Sensitivity and other checks To all parameters To nested CES technology Selection as Canada

Conclusion Skill-biased migration brings global efficiency gains But important distributional consequences: ◮ Positive effects in the receiving countries ◮ Losses in many sending countries

Thank you! Costanza Biavaschi costanza.biavaschi@ntnu.no

APPENDIX

The model - (links to) equations Labor market Consumer’s problem Firms Trade Remittances

Competitive labor markets Traditional sector: low-skilled only Q T i = A T i L T i Tradables/non-tradables 3 skill levels: low-, medium, and high-skilled � � σs σs − 1 σs − 1 σs − 1 Q M = A M α L + (1 − α L i − α H + α H σs − 1 i ( L i ) i ) ( M i ) i ( H i ) σs σs σs i i

Competitive labor markets Traditional sector: low-skilled only Q T i = A T i L T i Tradables/non-tradables 3 skill levels: low-, medium, and high-skilled � � σs σs − 1 σs − 1 σs − 1 Q M = A M α L + (1 − α L i − α H + α H σs − 1 i ( L i ) i ) ( M i ) i ( H i ) σs σs σs i i Immigrants and natives imperfect substitutes . Example for high skilled: � � σn σn − 1 . σn − 1 σn − 1 (1 − α F i )( H N + α F i ( H F H i = i ) i ) σn σn

Competitive labor markets Wages: � ( α L i ) σ s ( W L i ) 1 − σ s + W i = i ) 1 − σ s � 1 i ) 1 − σ s + ( α H 1 − σs . (1 − α L i − α H i ) σ s ( W M i ) σ s ( W H

Consumer’s problem Non-homothetic preferences 1 − β T � � � � θ { T i ,x ij ( k ) ,y i ( k ) } β T ( T i ) µ + θ − 1 θ − 1 θ − 1 max (1 − β )( Y i ) + β ( X i ) θ θ subject to: T i + P Y i Y i + P X i X i = w i ,   ǫ �� N Y � � N X ǫ ǫ − 1 � J ǫ − 1 j ǫ − 1 i ǫ − 1   ǫ dk ǫ dk X i = ( x ij ( k )) , Y i = ( y i ( k )) . 0 0 j =1

Indirect utility and price indices � β T µ � µ + (1 − β T ) w i − T i 1 − µ U i = β T 1 − β T P i . P i where P i is the ideal price index in country i , � � 1 − θ � (1 − β ) θ � � 1 − θ + β θ � 1 1 − θ , P Y P X P i = i i   1 � N X 1 − ǫ � J j   with: P X ( p ij ( k )) 1 − ǫ dk = , i 0 j =1 �� N Y � 1 1 − ǫ i and P Y ( p i ( k )) 1 − ǫ dk = . i 0

Positive for low skilled - really? Global 4 Change in real wages in % 2 0 -2 -4 WORLD OECD NON-OECD Real wages - baseline Real wages - current vs zero migration Real wages - current vs zero migration, no market size, perfect substitutability back

Extension: change nr of high-skilled migrants only Idea: reduce number of high-skilled migrants only ..until the skill-bias is eliminated Advantage: consistent with policy Problem: Change scale and selectivity back

Extension: change nr of high-skilled migrants only

Extension: remittances So far: every migrant remits a fixed amount , hence skill-biased migraton leaves remittances unaffected.

Extension: remittances So far: every migrant remits a fixed amount , hence skill-biased migraton leaves remittances unaffected. Three cases: ◮ Every migrants remits a fixed amount (minimalist) ◮ Every migrant remits a fixed share of his/her income (maximalist) ◮ Combination of the two cases (intermediate) Remittances are paid as a lump-sum transfer to non-migrants at origin

Non-OECD countries Change in welfare in % -6 -4 -2 0 2 Haiti Jamaica Albania Morocco El Salvador Zimbabwe Extension: remittances GHANA GHANA GHANA GHANA GHANA GHANA Welfare effect with remittances - maximalist Welfare effect with remittances - intermediate Welfare effect with remittances - minimalist Welfare effect - baseline Senegal Uruguay Moldova SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA Tunisia Philippines MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE Zambia Colombia Romania Vietnam INDIA INDIA INDIA INDIA INDIA INDIA Bangladesh BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL CHINA CHINA CHINA CHINA CHINA CHINA Mexico Ukraine Chile Bulgaria

OECD countries Change in welfare in % 0 1 2 3 4 Iceland Germany Estonia Italy FINLAND FINLAND FINLAND Austria Extension: remittances Belgium Welfare effect with remittances - maximalist Welfare effect with remittances - intermediate Welfare effect with remittances - minimalist Welfare effect - baseline Slovenia Denmark Spain Greece France Portugal Norway Sweden Netherlands New Zealand Ireland United Kingdom Switzerland UNITED STATES UNITED STATES UNITED STATES Australia Luxembourg Israel CANADA CANADA CANADA

Extension: remittances Global .6 Change in welfare in % .4 .2 0 -.2 -.4 WORLD OECD NON-OECD Welfare effect - baseline Welfare effect with remittances - minimalist Welfare effect with remittances - intermediate Welfare effect with remittances - maximalist

Extension: Brain gain effect Idea : migration creates incentives to invest in education Theory: Mountford (1997), Stark et al (1998), Beine et al (2001) Evidence: Beine et al (2008), Batista et al (2013), Shrestha (2015) � � � sh E − sh E � sh S = sh S 1 + σ b sh E

Non-OECD countries Change in welfare in % -6 -4 -2 0 2 Haiti Jamaica Albania Morocco Extension: Brain gain effect El Salvador Zimbabwe GHANA GHANA GHANA GHANA GHANA GHANA Welfare effect with brain gain - maximalist Welfare effect with brain gain - intermediate Welfare effect with brain gain - minimalist Welfare effect - baseline Senegal Uruguay Moldova SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA Tunisia Philippines MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE Zambia Colombia Romania Vietnam INDIA INDIA INDIA INDIA INDIA INDIA Bangladesh BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL CHINA CHINA CHINA CHINA CHINA CHINA Mexico Ukraine Chile Bulgaria

OECD countries Change in welfare in % 1.5 -.5 .5 0 1 2 Iceland Germany Estonia Italy Extension: Brain gain effect FINLAND FINLAND FINLAND Austria Belgium Welfare effect with brain gain - maximalist Welfare effect with brain gain - intermediate Welfare effect with brain gain - minimalist Welfare effect - baseline Slovenia Denmark Spain Greece France Portugal Norway Sweden Netherlands New Zealand Ireland United Kingdom Switzerland UNITED STATES UNITED STATES UNITED STATES Australia Luxembourg Israel CANADA CANADA CANADA

Extension: Brain gain effect 1 Change in welfare in % .5 0 -.5 WORLD OECD NON-OECD Welfare effect - baseline Welfare effect with brain gain - minimalist Welfare effect with brain gain - intermdiate Welfare effect with brain gain - maximalist

TFP Externalities Idea: TFP increases in the average level of human capital Theory: Lucas (1988) � � σ a H i A i = a i , H i + M i + L i

Non-OECD countries Change in welfare in % -60 -40 -20 0 Haiti Jamaica Albania Morocco El Salvador Welfare effect with Lucas externality on TFP - maximalist Welfare effect with Lucas externality on TFP - intermediate Welfare effect with Lucas externality on TFP - minimalist Welfare effect - baseline Zimbabwe GHANA GHANA GHANA GHANA GHANA GHANA TFP Externalities Senegal Uruguay Moldova SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA Tunisia Philippines MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE Zambia Colombia Romania Vietnam INDIA INDIA INDIA INDIA INDIA INDIA Bangladesh BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL CHINA CHINA CHINA CHINA CHINA CHINA Mexico Ukraine Chile Bulgaria

OECD countries Change in welfare in % 10 15 -5 0 5 Iceland Germany Estonia Italy FINLAND FINLAND FINLAND Welfare effect with Lucas externality on TFP - maximalist Welfare effect with Lucas externality on TFP - intermediate Welfare effect with Lucas externality on TFP - minimalist Welfare effect - baseline Austria Belgium TFP Externalities Slovenia Denmark Spain Greece France Portugal Norway Sweden Netherlands New Zealand Ireland United Kingdom Switzerland UNITED STATES UNITED STATES UNITED STATES Australia Luxembourg Israel CANADA CANADA CANADA

TFP Externalities 4 Change in welfare in % 2 0 -2 -4 WORLD OECD NON-OECD Welfare effect - baseline Welfare effect with Lucas externality on TFP - minimalist Welfare effect with Lucas externality on TFP - intermediate Welfare effect with Lucas externality on TFP - maximalist

Extension: Network effects in trade Immigrants foster trade with their home countries by reducing trade costs and demanding home-country-specific goods. Trade costs now: � � σ t H ij τ ij = ¯ τ ij ) H i j + M ij + L ij

Non-OECD countries Change in welfare in % -6 -4 -2 0 2 Haiti Jamaica Albania Morocco El Salvador Zimbabwe Extension: networks GHANA GHANA GHANA GHANA GHANA GHANA Senegal Uruguay Moldova Welfare effect - maximalist Welfare effect - intermediate Welfare effect - baseline SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA SOUTH AFRICA Tunisia Philippines MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE MOZAMBIQUE Zambia Colombia Romania Vietnam INDIA INDIA INDIA INDIA INDIA INDIA Bangladesh BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL BRAZIL CHINA CHINA CHINA CHINA CHINA CHINA Mexico Ukraine Chile Bulgaria

OECD countries Change in welfare in % -1 0 1 2 Iceland Germany Estonia Italy FINLAND FINLAND FINLAND Austria Extension: networks Belgium Slovenia Denmark Welfare effect - maximalist Welfare effect - intermediate Welfare effect - baseline Spain Greece France Portugal Norway Sweden Netherlands New Zealand Ireland United Kingdom Switzerland UNITED STATES UNITED STATES UNITED STATES Australia Luxembourg Israel CANADA CANADA CANADA

Extension: networks Global .8 Change in welfare in % .6 .4 .2 0 -.2 WORLD OECD NON-OECD Welfare effect - baseline Welfare effect - intermediate Welfare effect - maximalist

Downskilling Idea: not all high-skilled immigrants work in high-skilled jobs We re-calculate the share of high-skilled based on occupational distributions

OECD countries Change in welfare in % 1.5 -.5 .5 0 1 2 Iceland Germany Estonia Italy FINLAND FINLAND FINLAND Austria Belgium Slovenia Downskilling Denmark Welfare effect - downskilling Welfare effect - baseline Spain Greece France Portugal Norway Sweden Netherlands New Zealand Ireland United Kingdom Switzerland UNITED STATES UNITED STATES UNITED STATES Australia Luxembourg Israel CANADA CANADA CANADA