Skip to main content

Table 6 Estimation results from IV estimation in the AFT model

From: The productivity of return migrants: the case of China’s “Sea Turtles”

  25% longer imcomplete durations   Tobit
  OLS IV_LP IV_Probit IV_WR a   OLS IV_LP IV_Probit
  Coef. Coef. Coef. Coef.   Coef. Coef. Coef.
  (Std.) (Std.) (Std.) (Std.)   (Std.) (Std.) (Std.)
Foreign 0.12    0.42*   0.05   
  (0.07)    (0.19)   (0.07)   
Fitted_LP   0.31*      0.38*  
   (0.15)      (0.16)  
Fitted_Probit    0.36*      0.35*
    (0.16)      (0.16)
Experience -0.01* -0.01* -0.01* -0.02**   -0.02*** -0.02*** -0.02***
  (0.00) (0.00) (0.00) (0.01)   (0.00) (0.00) (0.00)
Offshore 0.25** 0.09 0.10 0.08   0.28** 0.00 0.04
  (0.09) (0.11) (0.11) (0.12)   (0.09) (0.11) (0.11)
JointVenture 0.18* 0.04 0.06 0.04   0.23* 0.00 0.04
  (0.08) (0.10) (0.10) (0.10)   (0.09) (0.10) (0.10)
Syndicate 0.04 0.03 0.01 0.00   0.15*** 0.12** 0.11*
  (0.04) (0.04) (0.04) (0.04)   (0.05) (0.04) (0.04)
TotalFund 0.01 0.01 0.01 0.01   0.02*** 0.02*** 0.02***
  (0.00) (0.01) (0.01) (0.01)   (0.01) (0.01) (0.01)
ExitRatio -0.04 -0.04 -0.04 -0.05   -0.00 0.00 0.00
  (0.03) (0.03) (0.03) (0.03)   (0.03) (0.03) (0.03)
Tenure1 -0.02*** -0.02*** -0.02*** -0.02***   -0.03*** -0.03*** -0.03***
  (0.01) (0.01) (0.01) (0.01)   (0.01) (0.01) (0.01)
Roundsize -0.01 0.00 -0.00 -0.00   0.00 0.00 -0.00
  (0.01) (0.01) (0.01) (0.01)   (0.01) (0.01) (0.01)
Tenure2 -0.11*** -0.11*** -0.10*** -0.10***   -0.18*** -0.18*** -0.17***
  (0.01) (0.01) (0.01) (0.01)   (0.01) (0.01) (0.01)
Development -0.27*** -0.25** -0.22** -0.21**   -0.08 -0.08 -0.08
  (0.08) (0.08) (0.08) (0.08)   (0.08) (0.08) (0.08)
Expansion -0.20* -0.24* -0.24* -0.22*   0.18 0.15 0.12
  (0.10) (0.10) (0.10) (0.10)   (0.11) (0.10) (0.10)
Late Stage -0.15 -0.17 -0.19 -0.17   0.44* 0.51** 0.45*
  (0.17) (0.17) (0.17) (0.17)   (0.20) (0.19) (0.19)
N 728 621 602 602   728 621 602
  1. * p < 0.05, ** p < 0.01, *** p < 0.001.
  2. Industry, region and year fixed effects are included in all regressions.
  3. The coefficients (standard errors) for linear probability and probit regressions in the first stage are 0.011. (0.004) and 0.028 (0.014), respectively.
  4. aIV_WR is a procedure Wooldridge proposed (see Wooldridge 2002, pp.623), which is to do a probit in the first stage, and then to estimate the empirical model using the fitted value as an IV. This approach yields more efficient estimates under certain conditions.