Skip to main content

Table 2 Static flow model estimates with alternative measurements for the welfare state

From: EU enlargement and the race to the bottom of welfare states

Explanatory variables Dependent variable: Δm ij,t  = annual change in the diaspora ratio
  (6) (7)
Destination country variables X j,t-1
Annual nominal net wage in 10,000 Euro in j 5.108*** 4.391*
  (1.610) (2.467)
Unemployment rate in % in j -0.401*** -0.420***
  (0.098) (0.115)
Population in destination country j in mio. persons 0.038* 0.039**
  (0.020) (0.019)
Southern European country (0/1) 9.272*** 9.305***
  (1.722) (1.704)
Regime dummy (restricted freedom of m.) in j (0/1) -11.059 -9.450
  (7.794) (8.401)
Net replacement rate in % in j 2.770  
  (4.860)  
Annual replacement income in 10,000 Euro in j   1.546
   (2.971)
Source-destination (network) effects N ij,t-1
Diaspora stock in j in mio. persons 0.508 0.710
  (2.842) (2.844)
Education level (ISCED) in the age-specific diaspora 2.964 3.043
  (1.845) (1.950)
Proportion of females in the age-specific diaspora -1.532 -1.592
  (3.130) (3.241)
Proportion of employed in the age-specific diaspora -3.480 -3.716
  (4.150) (4.055)
Source specific factors S i,t-1
Education level (ISCED) in the source cohort -3.515 -3.584
  (4.237) (4.327)
Proportion of employed in the source cohort -18.296** -18.679**
  (8.632) (8.133)
Interaction of push-pull factors and the migration regime I ij,t-1 a
Annual nominal net wage x regime dummy -4.129** -3.200
  (1.606) (2.531)
Unemployment rate x regime dummy 0.505*** 0.537***
  (0.145) (0.149)
Diaspora stock in j in mio. persons x regime dummy -1.393 -1.595
  (3.860) (3.891)
Net replacement rate x regime dummy -2.668  
  (4.487)  
Annual replacement income x regime dummy   -1.925
   (2.852)
Proportion of unemployed in the source cohort x -4.262 -3.813
regime dummy (16.629) (16.311)
Source cohort dummies Yes Yes
R-squared 0.475 0.475
Number of observations 432 432
  1. *p < 0.10, **p < 0.05, ***p < 0.01. Standard errors that are clustered by destination country and time periods are shown in parentheses. I use the STATA routine provided by Cameron et al. ([2011]) to control for two-way clustered error terms. In all regressions, time dummies and a constant are included. aFurther variables are included: see Additional file 1 for a table with all covariates. All explanatory variables are lagged by one year except for the migration regime dummy.