Instrument weakness-robust inferenc

Instrument weakness-robust inference can be achieved using the Anderson-Rubin (AR) (1949) statistic. The online Appendix discusses estimation of this statistic as well as the construction of weakness-robust confidence sets to assess the significance of each individual endogenous regressor. As shown in Table 9 , the AR-statistic strongly rejects the hypothesis that the relationship variables jointly equal zero. The AR confidence sets in the online Appendix show that each relationship indicator is individually significant. In summary, although reliable point estimates for relation- ship intensity are difficult to obtain using IV, the relation- ship effects are statistically significant even under weak- instrument robust inference.

結果 (繁體中文) 1:

可以使用安德森 - 魯賓（AR）（1949）統計量來實現儀器弱點魯棒的推斷。在線附錄討論這個統計的估計值，以及疲軟，強勁的信心套建築評估每個人的內源性回歸的意義。如表9中所示，AR-統計強烈反對的假設，即關係變量共同等於零。該AR信心套在網上的附錄顯示，每個關係的指標是分別顯著。綜上所述，雖然relation-船舶強度可靠的點估計很難獲得使用四，relation-船效果，即使在weak-儀強大的推論統計顯著。

結果 (繁體中文) 2:

CO TO ISLANDS OFFER LITTLE-KNOWN RETREAT

結果 (繁體中文) 3:

利用Anderson-Rubin（AR）（1949）統計量可以實現儀器的弱魯棒推斷。線上附錄討論了這一統計的估計，以及弱穩健置信集的構造，以評估每個個體內生回歸的重要性。如錶9所示，AR統計量強烈反對關係變數共同等於零的假設。線上附錄中的AR置信集表明，每一個關係名額都是單獨顯著的。綜上所述，雖然使用IV很難獲得關係强度的可靠點估計，但即使在弱儀器穩健推斷下，關係效應也具有統計顯著性。