Labor Market and Fiscal Impacts of International Migration: Lectures 3

1
Labor Market and Fiscal Impacts of International
Migration Lectures 3 4

• Gordon H. Hanson
• UCSD and NBER

2
Introduction

• Last three lectures
• What explains small scale, positive selection,
  and positive sorting in international migration
• Illegal migration
• Today
• What impact does international migration have on
  the labor markets of sending and receiving
  countries?
• Do changes in national labor supply affect wages?
• What are implications of global labor flows on
  fiscal accounts in sending countries for
  migrants?
• Does outflow of labor increase net tax burden on
  those that remain?

3
Rise in foreign born as share of OECD population
4
  1995 2000 2005 ?1995-2005
Australia 23.0 23.0 23.8 0.8
Austria 10.5 13.5
Belgium 9.7 10.3 12.1 2.4
Canada 16.6 17.4 19.1 2.5
Czech Republic 4.2 5.1 0.9
Denmark 4.8 5.8 6.5 1.7
Finland 2.0 2.6 3.4 1.4
Germany 11.5 12.5
Greece 10.3
Hungary 2.8 2.9 3.3 0.5
Ireland 6.9 8.7 11.0 4.1
Netherlands 9.1 10.1 10.6 1.5
New Zealand 16.2 17.2 19.4 3.2
Norway 5.5 6.8 8.2 2.7
Portugal 5.4 5.1 6.3 0.9
Slovak Republic 2.5 3.9
Spain 5.3
Sweden 10.5 11.3 12.4 1.9
Switzerland 21.4 21.9 23.8 2.4
United Kingdom 6.9 7.9 9.7 2.8
United States 9.3 11.0 12.9 3.6
5
Rising share of immigrants from low-income
countries in the OECD
Low Income Sending Region 1990 2000 Change
Mexico, Central America, Caribbean 0.149 0.202 0.053
Southeast Asia 0.086 0.102 0.016
Eastern Europe 0.057 0.099 0.042
Middle East 0.062 0.063 0.001
South Asia 0.041 0.052 0.011
North Africa 0.050 0.044 -0.006
South America 0.031 0.041 0.010
Central, Southern Africa 0.029 0.036 0.007
Former Soviet Union 0.031 0.029 -0.002
Total 0.540 0.672 0.132
High Income Sending Region
Western Europe 0.355 0.244 -0.111
Asia, Oceania 0.065 0.055 -0.010
North America 0.040 0.029 -0.011
Total 0.460 0.328 -0.132
6
Educational composition of immigration
Share in adult immigrant pop. Share in adult immigrant pop. Share in adult immigrant pop. Share in adult resident pop. Share in adult resident pop. Share in adult resident pop.
    primary education secondry education tertiary education   primary education secondry education tertiary education
1990 EU 8 0.616 0.210 0.174 0.332 0.485 0.183
Canada, US 0.388 0.180 0.433 0.118 0.486 0.397
Australia, N. Zealand 0.303 0.322 0.375 0.311 0.391 0.298
2000 EU 8 0.510 0.240 0.250 0.233 0.541 0.226
Canada, US 0.367 0.181 0.453 0.060 0.427 0.513
  Australia, N. Zealand 0.285 0.267 0.448   0.248 0.425 0.327
7
National emigration rates vary and are persistent
8
Theory

• Consider migration from a low wage country (eg,
  Mexico) to a high wage country (eg, the US)
• How would migration effect wages and national
  income in the two countries?
• Assumptions
• One good
• Two factors of production, labor and capital
• Only labor is mobile between countries
• All factors are fully employed (can relax this)
• Wage equals marginal revenue product

9
Solving for General Equilibrium

• Putting the elements of the model together
• Conditions for equilibrium in the labor market
• Wage equals marginal revenue product of labor in
  US
• WUS PUSMPLUS
• Wage equals marginal revenue product of labor in
  Mexico
• WMX PMXMPLMX
• World labor supply
• LUS LMX L

10
Solving for General Equilibrium

• Conditions for equilibrium in markets for capital
• Rental price of capital revenue MP of capital
  
• US RUSPUSMPKUS
• MX RMX PMXMPKMX
• We assume capital is immobile between countries
  (or whatever capital will move already has)
• As a result, equilibrium in K markets holds by
  assumption and remains in the background of the
  model
• All action comes from competing national demands
  for labor

11
Demand for Labor in the US
W
PUSMPLUS
LUS
12
Demand for Labor in Mexico
W
PMXMPLMX
LMX
13
Global Supply of Labor
W
? LMX
LUS ?
L
14
Labor Market Equilibrium with Labor Mobility
W
1
W1
PUSMPLUS
PMXMPLMX
? LMX
LUS ?
L1US
L1MX
15
Labor Market Equilibrium without Labor Mobility
W
WUS
WMX
PMXMPLMX
PUSMPLUS
? LMX
LUS ?
L0MX
L0US
16
Labor Market Equilibrium with Labor Mobility
W
WUS
W1
WMX
PMXMPLMX
PUSMPLUS
? LMX
LUS ?
L0MX
L0US
Immigrants
17
Labor Market Equilibrium with Labor Mobility
W
WUS
A
B
W1
C
F
D
WMX
PMXMPLMX
E
PUSMPLUS
? LMX
LUS ?
L0MX
L0US
Immigrants
18
Wage and Income Effects of Migration

• US natives
• Loss in labor income A
• Gain in capital income B
• Gain in GNP B
• Gain in GDP BCDE
• Mexico natives
• Gain in labor income (migrants) CD
• Gain in labor income (non-migrants) F
• Loss in capital income DF
• Gain in GNP C
• Loss in GDP DE
• Gain in global GNP BC

19
Trends in migrant share of population, N. America
20
Impact of Emigration on Wages

• Does the exodus of labor due to emigration drive
  up wages in sending countries?
• Estimated elasticity of wages with respect to
  emigration for Mexico
• Using natl data, 0.4-0.6 (Mishra, Aydemir
  Borjas)
• Using regional data, 0.7 (Hanson)
• Problems in estimating the wage elasticity of
  emigration (all of which would appear to bias
  estimate toward zero)
• Estimate is reduced form, net of effect of
  capital adjustment
• Emigration may be negatively correlated with
  local wage shocks

21
Impact of Emigration on Wages

• Estimation framework
• Demand for labor (L)
• Demand for leisure (H)
• Labor supply constraint
• Solving for wages (W) as a function of emigration
  (M)

22
Impact of Emigration on Wages

• Estimation framework
• Specification
• The wage for a given education (E) and experience
  (X) cell at time t is a function of the
  emigration shock and other shocks
• E, X, T dummies control for time-invariant shocks
  specific to each education/experience group and
  year-specific shocks for each education group and
  each experience group
• Identification is based on variation in the
  emigration shock within education-experience
  groups over time

23
Emigrants as a share of male population by
education group, Mexico 1960-2000
24
Labor impacts in receiving countries

• There is an enormous literature on the impact of
  foreign labor inflows on the US labor market
• Studies using national-level data find that
  immigration reduces wages for US native workers,
  with strongest effects being felt by low-skilled
  workers
• Studies using city or state-level data find that
  immigrations wage impacts are small
• What are flaws in two approaches to studying
  labor market impacts of immigration? What is
  true impact of foreign labor inflows on labor
  markets?

25
Immigrant share of US workforce, by education
26
US Immigrant supply shocks, by skill group
27
Immigrant share of Canadian workforce, by
education
28
National-level regression (Borjas, 2003)

• Let yijt be the mean value of a particular labor
  market outcome for native with education s,
  experience x, in year t. Stack the data across
  skill groups and calendar years and estimate
• yext ? pext E X T (E T) (X T)
  (E X) ?ext,
• E are fixed effects indicating educational
  attainment X are fixed effects indicating work
  experience T are fixed effects indicating
  calendar year p is the immigrant share in the
  skill group.
• Regressions weighted by sample size of
  education-experience-year cell. Standard errors
  clustered by education-experience.
• Problems with this approach
• Are there other shocks that could be correlated
  with immigration inflows (eg, skill biased
  technological change)?
• What is counterfactual? Could immigration cause
  changes in technology (which would have changed
  fixed effects)?

29
National-level results for US
Log weekly earnings Employment rate (log odds) Incarceration rate (log odds)
A. Least squares
Blacks -0.561 -0.489 0.100
(0.167) (0.092) (0.040)
Whites -0.669 -0.206 0.010
(0.250) (0.079) (0.015)
B. IV
Blacks -0.519 -0.488 0.105
(0.172) (0.098) (0.038)
Whites -0.640 -0.180 0.003
(0.255) (0.088) (0.016)
Multiply these coefficients by 0.73 to get a
numerically sensible elasticity dy/d( change in
supply)
30
US Simulation results, log weekly wage
1. Actual change 2. Predicted impact 3. due to immigration
Blacks
High school dropouts -0.140 -0.083 59.5
High school graduates -0.085 -0.032 37.4
Some college 0.008 -0.033 --
College graduates 0.127 -0.044 --
All persons -0.060 -0.036 60.0
Whites
High school dropouts -0.206 -0.086 41.7
High school graduates -0.146 -0.033 22.8
Some college 0.050 -0.035 --
College graduates 0.102 -0.046 --
All persons -0.060 -0.037 61.7
31
Has immigration really changed the US wage
structure?
32
Correlation in native wages immigrant inflows
for US cities (Card, 2005)
33
Assessing the literature

• National level data find negative correlation
  between native wages and immigrant inflows by
  skill group
• Are all relevant shocks controlled for?
• Could immigration affect technology?
• Area studies find a zero correlation between
  wages for low-skilled natives and immigration
  inflows
• Isnt the size of the local immigrant population
  endogenous (with more immigrants in locations
  with higher wages)?
• How does one instrument for local immigrant stock?

34
Resolving difference in results at national
local level

• Possible explanations
• Local technology is endogenous to labor supply
  (Acemoglu, 1998 Lewis, 2005)
• Possible but insufficient data to evaluate to
  date
• Internal native migration hides wage effects at
  local level (Card, 2001 Borjas, 2006)
• Conflicting results in the literature
• Explanations rejected by the data
• Immigrant and natives are imperfect substitutes
  (Ottaviano Peri, 2005)
• Plausible but data are inconsistent with this
  hypothesis in the US
• Sectoral output mix adjusts to absorb immigrants
  without wage effects as in HO model (Lewis, 2005
  Hanson Slaughter, 2001)
• Data reject this hypothesis

35
Emigration and Remittances
Workers' Remittances as of GDP Workers' Remittances as of GDP Workers' Remittances as of GDP Workers' Remittances as of GDP
Region 1992 2000 2005
East Asia Pacific 0.6 1.0 1.5
Europe Central Asia 1.4 1.4
Latin America Caribbean 0.7 1.0 2.0
Middle East North Africa 8.3 3.1 4.1
South Asia 1.8 2.9 3.5
Sub-Saharan Africa 0.8 1.5 1.6
36
Emigration and Remittances

• Emigrants share income gains with family members
  through remittances
• Fulfillment of migration contract (Lucas Stark)
  and/or consumption smoothing across borders
  (Rosenzweig Stark)
• Positive income shock in emigrants country is
  associated with larger remittances to sending
  country (Yang)
• Remittances support both consumption and
  investment spending (Yang) and may deepen
  financial markets (Woodruff et al.)
• But there is no reason to believe primary
  motivation of remittances is investment in the
  sending country
• Impediments to remittances
• Service fees on money transfers average 11 and
  elasticity of remittances w.r.t. services fees is
  -0.15 (Freund Spatafora)

37
Effects of emigration on public finances

• The outflow of labor means that sending countries
  lose access to future tax payments of emigrants
• But they also shed the obligation to provide
  departing workers with income transfers
• What is net effect of tax loss and spending
  savings?
• Presumption is that the net effect is negative
• Most emigrants are relatively highly educated
• Emigrants receive many of their lifetime
  transfers prior to their departure in the form of
  public education and child healthcare

38
Fiscal impacts of emigration in India (Desai et
al.)

• In 2000, individuals with tertiary education made
  up 60.5 of Indian emigrants but just 4.5 of
  Indias total population
• Between 1990 and 2000, emigration rate for
  tertiary educated rose from 2.8 to 4.3 (but
  only 0.3 to 0.4 for total population
• US has 64.6 of Indias skilled emigrants (48.9
  of all Indian emigrants)
• Estimation procedure
• Produce counterfactual incomes income emigrants
  would have earned in India based on observed
  characteristics and returns to these
  characteristics in India (using Mincer wage
  regression)
• Calculate income tax losses by running
  counterfactual income through Indian income tax
  schedule (and calculate indirect tax losses)
• Calculate expenditure savings by identifying
  categories for which savings would exist eg,
  not interest payments and national defense
• Emigration to US cost India net tax contributions
  of 0.24 of GDP in 2000, partially offset by the
  tax take on remittances of 0.1 of GDP

39
Other Impacts of Emigration

• Does emigration affect international trade?
• Bilateral migration is positively correlated with
  bilateral trade, though causality is unclear
  (Gould, Head Reis)
• Countries with larger bilateral Chinese
  populations do trade more, especially in
  differentiated products (Rauch Trindade)
• Does emigration affect international knowledge
  flows?
• Chinese and Indian engineers in Silicon Valley
  have contributed to FDI and business formation in
  their home countries (Saxenian)
• Countries that send students to democratic
  countries for university have stronger democratic
  leanings later on (Spilimbergo)

40
What We Dont Know

• The impact of emigration on wages and net tax
  revenues
• Do results for Mexico and India generalize?
• The impact of skilled emigration on the stock of
  human capital
• What is the primary cause of skilled emigration?
• Does skilled emigration cause a brain drain or a
  brain gain?
• The relative contribution of labor productivity,
  returns to skill, and migration costs to
  migration flows
• Are migration costs large enough to explain small
  migration flows?
• Is there a skill bias in receiving country
  immigration policies?
• How do sending-country policies affect
  emigration?
• How does migration interact with trade and
  capital flows?
• Are trade and migration substitutes or
  complements?
• Do remittances help deepen financial markets?