IV in R - Revision history

Rb at 22:56, 7 August 2015

2015-08-07T22:56:38Z

Rb: /* Sargan test for instrument validity */

2015-08-07T22:56:11Z

‎Sargan test for instrument validity

Rb: /* Sargan test for instrument validity */

2015-08-07T22:55:51Z

‎Sargan test for instrument validity

Rb: /* Sargan test for instrument validity */

2015-08-07T22:55:12Z

‎Sargan test for instrument validity

Rb: /* Testing for exogeneity */

2015-08-07T22:53:27Z

‎Testing for exogeneity

Rb: /* Implementation in R */

2015-08-07T22:46:51Z

‎Implementation in R

Rb: /* Example */

2015-08-07T22:40:42Z

‎Example

Rb: /* Example */

2015-08-07T22:38:44Z

‎Example

Rb: /* Example */

2015-08-07T22:38:11Z

‎Example

Rb: /* Example */

2015-08-07T22:36:03Z

‎Example

← Older revision		Revision as of 22:56, 7 August 2015
Line 1:		Line 1:
−	~~This section is currently being written, especially the testing part.~~

	= Introduction =		= Introduction =

@@ Line 220: / Line 220: @@
       print(1-pchisq(Sargan_test,1))  # prints p-value
-We find that the p-value of this test is 0.5260 and hence we do not reject the null hypothesis of instrument validity. The p-value was obtained from a $\Chi^2$ distribution with one degree of freedom. That was one here as we had two instruments for one endogenous variable (2-1).
+We find that the p-value of this test is 0.5260 and hence we do not reject the null hypothesis of instrument validity. The p-value was obtained from a $\chi^2$ distribution with one degree of freedom. That was one here as we had two instruments for one endogenous variable (2-1).
 =Footnotes=
   <references />

@@ Line 211: / Line 211: @@
 The test is rather simple to implement. Take the residuals from the 2SLS regression <code>reg_iv1$residuals</code> and use them as the dependent variable in a new regression in which you regress them on all exogenous explanatory variables and all instruments.
-      Sargan_reg <- lm(<code>reg_iv1<code>$</code>residuals</code>~age+exper+expersq+fatheduc+motheduc,data=mydata)
+      Sargan_reg <- lm(<code>reg_iv1$residuals</code>~age+exper+expersq+fatheduc+motheduc,data=mydata)
       Sargan_reg_sm <- summary(Sargan_reg)

@@ Line 211: / Line 211: @@
 The test is rather simple to implement. Take the residuals from the 2SLS regression <code>reg_iv1$residuals</code> and use them as the dependent variable in a new regression in which you regress them on all exogenous explanatory variables and all instruments.
-      Sargan_reg <- lm(<code>reg_iv1$residuals</code>~age+exper+expersq+fatheduc+motheduc,data=mydata)
+      Sargan_reg <- lm(<code>reg_iv1<code>$</code>residuals</code>~age+exper+expersq+fatheduc+motheduc,data=mydata)
       Sargan_reg_sm <- summary(Sargan_reg)
 If the instruments are valid (null hypothesis), they should be uncorrelated to these residuals and hence we apply the following $\chi^2$ test. We use the $R^2$ of this regression and calculate $n*R^2$
-      Sargan_test <- Sargan_reg_sm$r.squared*nrow(mydata)
+      Sargan_test <- Sargan_reg_sm<code>$</code>r.squared*nrow(mydata)
       print(Sargan_test)
       print(1-pchisq(Sargan_test,1))  # prints p-value
-We find that the p-value of this test is 0.5260 and hence we do not reject the null hypothesis of instrument validity. The p-value was obtained from a $|chi^2$ distribution with one degree of freedom. That was one here as we had two instruments for one endogenous variable (2-1).
+We find that the p-value of this test is 0.5260 and hence we do not reject the null hypothesis of instrument validity. The p-value was obtained from a $\Chi^2$ distribution with one degree of freedom. That was one here as we had two instruments for one endogenous variable (2-1).
 =Footnotes=
   <references />

@@ Line 195: / Line 195: @@
       print(HausWutest)
       Wald test
-      Model 1: lwage ~ educ + age + exper + expersq + first_stage$residuals
+      Model 1: lwage ~ educ + age + exper + expersq + first_stage<code>$</code>residuals
       Model 2: lwage ~ educ + age + exper + expersq
         Res.Df Df      F  Pr(>F)

← Older revision		Revision as of 22:46, 7 August 2015
Line 28:		Line 28:

	The R Package needed is the AER package that we already recommended for use in the context of estimating [[R_robust_se#Which_package_to_use\|robust standard errors]]. Included in that package is a function called <source enclose=none>ivreg</source> which we will use. We explain how to use it by walking through an example.		The R Package needed is the AER package that we already recommended for use in the context of estimating [[R_robust_se#Which_package_to_use\|robust standard errors]]. Included in that package is a function called <source enclose=none>ivreg</source> which we will use. We explain how to use it by walking through an example.
		+
		+	If you use IV a lot in your work, you may well want to pack all of the following into one convenient function (just as Alan Fernihough has done here [https://diffuseprior.wordpress.com/2012/05/03/an-ivreg2-function-for-r/]. But if you are applying IV for the first time it is actually very instructive to go through some of the steps in a bit more detail. It is is also good to see that really there is not a lot of technical magic ... just a great idea!

	== Example ==		== Example ==

@@ Line 137: / Line 137: @@
       Multiple R-Squared: 0.1353,	Adjusted R-squared: 0.1272
       Wald test: 6.085 on 4 and 423 DF,  p-value: 9.085e-05
 = IV related Testing procedures =

@@ Line 118: / Line 118: @@
       Call:
       ivreg(formula = lwage ~ educ + age + exper + expersq | . - educ +
-          fatheduc, data = mydata)
+          fatheduc + motheduc, data = mydata)
       Residuals:
-          Min       1Q   Median       3Q      Max
+         Min      1Q  Median      3Q     Max
-      -3.09354 -0.32798  0.05094  0.37402  2.35375
+      -3.1017 -0.3216  0.0545  0.3685  2.3496
       Coefficients:
                     Estimate Std. Error t value Pr(>|t|)
-      (Intercept) -0.0513505  0.4936538  -0.104  0.91720
+      (Intercept)  0.0667186  0.4591101   0.145   0.8845
-      educ         0.0701490  0.0346051   2.027  0.04328 *
+      educ         0.0609945  0.0315798   1.931   0.0541 .
-      age         -0.0002287  0.0049140  -0.047  0.96290
+      age         -0.0003542  0.0049318  -0.072   0.9428
-      exper        0.0436778  0.0134180   3.255  0.00122 **
+      exper        0.0441960  0.0134524   3.285   0.0011 **
-      expersq     -0.0008790  0.0004064  -2.163  0.03111 *
+      expersq     -0.0008947  0.0004075  -2.196   0.0287 *
       ---
       Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
-−
-     Residual standard error: 0.6727 on 423 degrees of freedom
+Residual standard error: 0.6756 on 423 degrees of freedom
-     Multiple R-Squared: 0.143,	Adjusted R-squared: 0.1349
+Multiple R-Squared: 0.1353,	Adjusted R-squared: 0.1272
-     Wald test: 6.225 on 4 and 423 DF,  p-value: 7.106e-05
+Wald test: 6.085 on 4 and 423 DF,  p-value: 9.085e-05
 = IV related Testing procedures =

@@ Line 35: / Line 35: @@
 Let's first import the data
-      setwd("T:/ECLR/R/FirstSteps")              # This sets the working directory
+      setwd("YOUR/DIRECTORY/PATH")              # This sets the working directory
       mydata <- read.csv("mroz.csv", na.strings = ".")  # Opens mroz.csv from working directory