Skip to main content

Table 3 Generalized linear instrumental variable model and first stage equation

From: Overconfidence and the adoption of robo-advice: why overconfident investors drive the expansion of automated financial advice

 

(1) IRT OC

(2) Robo-advice

No prediction S&P500

− 0.53***

 
 

(0.06)

 

Fitted IRT overconfidence

 

0.75**

  

(0.24)

Willingness to take risk

0.15***

0.08

 

(0.01)

(0.06)

Female

− 0.17***

0.07

 

(0.04)

(0.10)

Age categories

  

25–34

− 0.04

− 0.07

 

(0.12)

(0.19)

35–44

− 0.09

− 0.32

 

(0.12)

(0.21)

45–54

− 0.14

− 0.72**

 

(0.12)

(0.23)

55–64

− 0.08

− 0.79***

 

(0.12)

(0.23)

65+

− 0.10

− 1.19***

 

(0.12)

(0.26)

Bachelor or more

− 0.04

− 0.03

 

(0.04)

(0.09)

Household income

  

$50–$100 K

0.05

− 0.20

 

(0.06)

(0.11)

$100 K or more

0.16*

− 0.56***

 

(0.06)

(0.13)

Married

0.04

0.17

 

(0.05)

(0.10)

Ethnic/racial minority

0.14**

0.11

 

(0.05)

(0.11)

Constant

− 0.82***

− 1.11**

 

(0.14)

(0.43)

Observations

1,937

1,937

  1. Column (1) illustrates the first stage in Eq. 3. Robust standard errors in parentheses. Z-score of “No Prediction S&P 500” equals − 8.39. Column (2) is a simultaneous equation generalized probit model, replacing IRT OC with its estimated values from the first stage
  2. ***p < 0.001; **p < 0.01; *p < 0.05