“Paquetes”
library(stargazer)##
## Please cite as:## Hlavac, Marek (2018). stargazer: Well-Formatted Regression and Summary Statistics Tables.## R package version 5.2.2. https://CRAN.R-project.org/package=stargazerlibrary(lmtest)## Loading required package: zoo##
## Attaching package: 'zoo'## The following objects are masked from 'package:base':
##
## as.Date, as.Date.numericlibrary(AER)## Loading required package: car## Loading required package: carData## Loading required package: sandwich## Loading required package: survivallibrary(markdown)“Base de Datos”
library(haven)
mus06data <- read_dta("mus06data.dta")
View(mus06data)\(ldrugexp =\beta_{0} + \beta_{1}hi_empunion + \beta_{2}totchr + \beta_{3}age + \beta_{4}female + \beta_{5}blhisp + \beta_{6}linc + u \)
Donde:
Ldrugexp= Es el logaritmo del desembolso total en medicamentos recetados
Hi_empunion= Si la persona tiene seguro de salud patrocinado por el empleador o por el sindicato (variable binaria)
Totchr= Número de enfermedades crónicas
Age= Edad en años
Female= Mujer (variable binaria)
Blhisp= Si es negro o Hispano (variable binaria)
Linc= logaritmo natural de los ingresos anuales del hogar en miles de dolares
EstMCO <- lm(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc, data=mus06data )
summary(EstMCO)##
## Call:
## lm(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.3295 -0.6754 0.1516 0.8559 3.7343
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 5.861131 0.153184 38.262 < 2e-16 ***
## hi_empunion 0.073879 0.026109 2.830 0.00467 **
## totchr 0.440381 0.009573 46.002 < 2e-16 ***
## age -0.003529 0.001886 -1.871 0.06132 .
## female 0.057806 0.025163 2.297 0.02163 *
## blhisp -0.151307 0.033808 -4.475 7.71e-06 ***
## linc 0.010482 0.013952 0.751 0.45251
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.236 on 10082 degrees of freedom
## (302 observations deleted due to missingness)
## Multiple R-squared: 0.177, Adjusted R-squared: 0.1765
## F-statistic: 361.3 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(EstMCO, type = 'text', title = "Regresión por MCO", df = F, digits = 4)##
## Regresión por MCO
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion 0.0739***
## (0.0261)
##
## totchr 0.4404***
## (0.0096)
##
## age -0.0035*
## (0.0019)
##
## female 0.0578**
## (0.0252)
##
## blhisp -0.1513***
## (0.0338)
##
## linc 0.0105
## (0.0140)
##
## Constant 5.8611***
## (0.1532)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.1770
## Adjusted R2 0.1765
## Residual Std. Error 1.2360
## F Statistic 361.3167***
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01Interpretación:
El desembolso total en medicamentos recetados de una persona que tenga seguro de salud patrocinado por el empleador o por el sindicato es mayor en 7.4% que una persona que no tiene seguro de salud, y esta variable es significativa
Por cada enfermedad crónica que se padezca, el desembolso total en medicamentos recetados aumenta un 44%, y esta variable es significativa
Por cada año más que tenga la persona, el desembolso total en medicamentos recetados disminuye .4%, pero no es muy significativa esta variable
El desembolso total en medicamentos recetados de una mujer es 5.8% mayor que el de un hombre y es una variable medianamente significativa
El desembolso total en medicamentos recetados de una persona negra o hispana es 15.1% menor que una persona que no lo es, y es una variable significativa
Por cada unidad porcentual que aumente el ingreso anual del hogar, el desembolso total en medicamentos recetados aumenta .010%, pero es una variable no significativa
Cuando todas las variables dependientes valen 0, el desembolso total en medicamentos recetados es de 586.1% y es una variable significativa.
Cabe resaltar que cada interpretación es en promedio, ceteris paribus, es decir, que por cada variación porcentual que estemos presentando, todo lo demás permanecerá constante.
El modelo se ajusta en un 17.7% a los datos reales.
La medida de bondad de ajuste de R2 tiene un 17.6%, este intenta corregir la estimación excesiva cuando aumentan el número de efectos incluidos en el modelo.
la variable hi_empunion la trataremos como endógena porque tener dicho seguro complementario además del seguro casi universal de Medicare para las personas mayores, puede ser una variable de elección. A pesar de que la mayoría de los individuos en la muestra ya no trabajan, aquellos que esperaban altos gastos médicos futuros podrían haber tenido más probabilidades de elegir un trabajo que les proporcionará un seguro de salud complementario al jubilarse.
Probando el primer instrumento ssiratio (es la relación entre el ingreso de seguridad social de un individuo y el ingreso del individuo de todas las fuentes)
\(hi_empunion = \beta_{0} + \beta{1}ssiratio + \beta{2}totchr + \beta{3}age + \beta{4}female + \beta{5}blhisp + \beta{6}linc + v \)
inst1 <- lm(hi_empunion ~ ssiratio + totchr + age + female + blhisp + linc, data=mus06data )
summary(inst1)##
## Call:
## lm(formula = hi_empunion ~ ssiratio + totchr + age + female +
## blhisp + linc, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.7162 -0.3879 -0.2321 0.5119 2.5291
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.028981 0.057409 17.924 < 2e-16 ***
## ssiratio -0.191643 0.014129 -13.564 < 2e-16 ***
## totchr 0.012786 0.003622 3.530 0.000418 ***
## age -0.008632 0.000713 -12.107 < 2e-16 ***
## female -0.073450 0.009493 -7.737 1.11e-14 ***
## blhisp -0.062680 0.012769 -4.909 9.30e-07 ***
## linc 0.048394 0.005677 8.525 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4672 on 10082 degrees of freedom
## (302 observations deleted due to missingness)
## Multiple R-squared: 0.07605, Adjusted R-squared: 0.0755
## F-statistic: 138.3 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(inst1, type = 'text', title = "Prueba Instrumento 1", df = F, digits = 4)##
## Prueba Instrumento 1
## ===============================================
## Dependent variable:
## ---------------------------
## hi_empunion
## -----------------------------------------------
## ssiratio -0.1916***
## (0.0141)
##
## totchr 0.0128***
## (0.0036)
##
## age -0.0086***
## (0.0007)
##
## female -0.0734***
## (0.0095)
##
## blhisp -0.0627***
## (0.0128)
##
## linc 0.0484***
## (0.0057)
##
## Constant 1.0290***
## (0.0574)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0761
## Adjusted R2 0.0755
## Residual Std. Error 0.4672
## F Statistic 138.3165***
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01Probando el segundo instrumento lowincome (es un indicador cualitativo del estado de bajos ingresos) \(hi_empunion = \beta_{0} + \beta{1}lowincome + \beta{2}totchr + \beta{3}age + \beta{4}female + \beta{5}blhisp + \beta{6}linc + v \)
inst2 <- lm(hi_empunion ~ lowincome + totchr + age + female + blhisp + linc, data=mus06data )
summary(inst2)##
## Call:
## lm(formula = hi_empunion ~ lowincome + totchr + age + female +
## blhisp + linc, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.7028 -0.3910 -0.2473 0.5227 1.2557
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.9709851 0.0575518 16.871 < 2e-16 ***
## lowincome -0.0907429 0.0123111 -7.371 1.83e-13 ***
## totchr 0.0105194 0.0036408 2.889 0.00387 **
## age -0.0097099 0.0007115 -13.647 < 2e-16 ***
## female -0.0791551 0.0095412 -8.296 < 2e-16 ***
## blhisp -0.0668208 0.0128450 -5.202 2.01e-07 ***
## linc 0.0707178 0.0053621 13.189 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4702 on 10082 degrees of freedom
## (302 observations deleted due to missingness)
## Multiple R-squared: 0.06424, Adjusted R-squared: 0.06368
## F-statistic: 115.3 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(inst2, type = 'text', title = "Prueba Instrumento 2", df = F, digits = 4)##
## Prueba Instrumento 2
## ===============================================
## Dependent variable:
## ---------------------------
## hi_empunion
## -----------------------------------------------
## lowincome -0.0907***
## (0.0123)
##
## totchr 0.0105***
## (0.0036)
##
## age -0.0097***
## (0.0007)
##
## female -0.0792***
## (0.0095)
##
## blhisp -0.0668***
## (0.0128)
##
## linc 0.0707***
## (0.0054)
##
## Constant 0.9710***
## (0.0576)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0642
## Adjusted R2 0.0637
## Residual Std. Error 0.4702
## F Statistic 115.3484***
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01Probando el tercer instrumento multlc (indica si la empresa es un gran operador con múltiples localizaciones) \(hi_empunion = \beta_{0} + \beta{1}multlc + \beta{2}totchr + \beta{3}age + \beta{4}female + \beta{5}blhisp + \beta{6}linc + v \)
inst3 <- lm(hi_empunion ~ multlc + totchr + age + female + blhisp + linc, data=mus06data )
summary(inst3)##
## Call:
## lm(formula = hi_empunion ~ multlc + totchr + age + female + blhisp +
## linc, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.7715 -0.3822 -0.2594 0.5357 1.2940
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.9016900 0.0579164 15.569 < 2e-16 ***
## multlc 0.1487593 0.0200300 7.427 1.20e-13 ***
## totchr 0.0109104 0.0036417 2.996 0.00274 **
## age -0.0091799 0.0007186 -12.775 < 2e-16 ***
## female -0.0792221 0.0095406 -8.304 < 2e-16 ***
## blhisp -0.0741602 0.0128696 -5.762 8.53e-09 ***
## linc 0.0720981 0.0053257 13.538 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4702 on 10082 degrees of freedom
## (302 observations deleted due to missingness)
## Multiple R-squared: 0.06431, Adjusted R-squared: 0.06376
## F-statistic: 115.5 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(inst3, type = 'text', title = "Prueba Instrumento 3", df = F, digits = 4)##
## Prueba Instrumento 3
## ===============================================
## Dependent variable:
## ---------------------------
## hi_empunion
## -----------------------------------------------
## multlc 0.1488***
## (0.0200)
##
## totchr 0.0109***
## (0.0036)
##
## age -0.0092***
## (0.0007)
##
## female -0.0792***
## (0.0095)
##
## blhisp -0.0742***
## (0.0129)
##
## linc 0.0721***
## (0.0053)
##
## Constant 0.9017***
## (0.0579)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0643
## Adjusted R2 0.0638
## Residual Std. Error 0.4702
## F Statistic 115.4953***
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01Probando el cuarto instrumento firmsz (mide el tamaño de la fuerza laboral empleada de la empresa) \(hi_empunion = \beta_{0} + \beta{1}firmsz + \beta{2}totchr + \beta{3}age + \beta{4}female + \beta{5}blhisp + \beta{6}linc + v \)
inst4 <- lm(hi_empunion ~ firmsz + totchr + age + female + blhisp + linc, data=mus06data )
summary(inst4)##
## Call:
## lm(formula = hi_empunion ~ firmsz + totchr + age + female + blhisp +
## linc, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.8755 -0.3892 -0.2609 0.5308 1.3424
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.9481012 0.0576506 16.446 < 2e-16 ***
## firmsz 0.0068293 0.0021640 3.156 0.00160 **
## totchr 0.0102640 0.0036486 2.813 0.00492 **
## age -0.0099029 0.0007125 -13.898 < 2e-16 ***
## female -0.0807026 0.0095605 -8.441 < 2e-16 ***
## blhisp -0.0687961 0.0128742 -5.344 9.3e-08 ***
## linc 0.0784064 0.0052624 14.899 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4713 on 10082 degrees of freedom
## (302 observations deleted due to missingness)
## Multiple R-squared: 0.06012, Adjusted R-squared: 0.05956
## F-statistic: 107.5 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(inst4, type = 'text', title = "Prueba Instrumento 4", df = F, digits = 4)##
## Prueba Instrumento 4
## ===============================================
## Dependent variable:
## ---------------------------
## hi_empunion
## -----------------------------------------------
## firmsz 0.0068***
## (0.0022)
##
## totchr 0.0103***
## (0.0036)
##
## age -0.0099***
## (0.0007)
##
## female -0.0807***
## (0.0096)
##
## blhisp -0.0688***
## (0.0129)
##
## linc 0.0784***
## (0.0053)
##
## Constant 0.9481***
## (0.0577)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0601
## Adjusted R2 0.0596
## Residual Std. Error 0.4713
## F Statistic 107.4883***
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01Interpretación:
Dadas las 4 regresiones anteriores, para probar los instrumentos posibles para “Hi_empunion”, podemos observar que los 4 instrumentos son fuertes, ya que presentan un nivel de significancia de al menos un .05%. Sin embargo, ya corriendo cada regresión y observando que tan significantes pueden ser cada uno de ellos, destacamos los instrumentos ssiratio, lowincome y multlc en ese orden respectivamente.
El instrumento que consideramos mejor para la variable “Hi_empunion” (Si la persona tiene seguro de salud patrocinado por el empleador o por el sindicato), por su nivel de significancia es “ssiratio”, es decir, la relación entre el ingreso de seguridad social de un individuo y el ingreso del individuo de todas las fuentes. Intuitivamente, también lo consideramos como el mejor instrumento ya que el ingreso es un determinante para que el empleado considere contratar un seguro, ya que de no contar con un ingreso suficiente, puede que busque sindicalizarse para obtener dicho seguro.
todas <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc |totchr + age + female + blhisp + linc + ssiratio , data=mus06data)
summary(todas)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + ssiratio,
## data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.7616 -0.7529 0.1275 0.8959 4.0723
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.787170 0.255523 26.562 < 2e-16 ***
## hi_empunion -0.897591 0.207991 -4.316 1.61e-05 ***
## totchr 0.450266 0.010422 43.201 < 2e-16 ***
## age -0.013218 0.002876 -4.596 4.36e-06 ***
## female -0.020406 0.031552 -0.647 0.518
## blhisp -0.217424 0.038688 -5.620 1.96e-08 ***
## linc 0.087002 0.022022 3.951 7.85e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.318 on 10082 degrees of freedom
## Multiple R-Squared: 0.06395, Adjusted R-squared: 0.0634
## Wald test: 319.6 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(todas, type = 'text', title = "Mejor Instrumento", df = F, digits = 4)##
## Mejor Instrumento
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.8976***
## (0.2080)
##
## totchr 0.4503***
## (0.0104)
##
## age -0.0132***
## (0.0029)
##
## female -0.0204
## (0.0316)
##
## blhisp -0.2174***
## (0.0387)
##
## linc 0.0870***
## (0.0220)
##
## Constant 6.7872***
## (0.2555)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0640
## Adjusted R2 0.0634
## Residual Std. Error 1.3182
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01Interpretación:
El desembolso total en medicamentos recetados de una persona que tenga seguro de salud patrocinado por el empleador o por el sindicato es menor en 89.75% que una persona que no tiene seguro de salud.
Por cada enfermedad crónica que se padezca, el desembolso total en medicamentos recetados aumenta un 44%
Por cada año más que tenga la persona, el desembolso total en medicamentos recetados disminuye en 1.32%
El desembolso total en medicamentos recetados de una mujer es 2.04% menor que el de un hombre y, a diferencia de las demás variables, esta es una variable no significativa
El desembolso total en medicamentos recetados de una persona negra o hispana es 21.74% menor que una persona que no lo es
Por cada unidad porcentual que aumente el ingreso anual del hogar, el desembolso total en medicamentos recetados aumenta 8.7%,
Cuando todas las variables dependientes valen 0, el desembolso total en medicamentos recetados es de 678.71% y es una variable significativa.
Cabe resaltar que cada interpretación es en promedio, ceteris paribus, es decir, que por cada variación porcentual que estemos presentando, todo lo demás permanecerá constante.
Por variables instrumentales no tomaremos en cuenta la R cuadrada ya que tenemos menos precsión pero nos estamos acercando a los datos reales, es decr, vamos a tener menos eficiencia (porque los errores estandar aumentan), pero tendremos más consistencia, lo cual es mejor.
todas <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc |totchr + age + female + blhisp + linc + ssiratio , data=mus06data)
summary(todas,diagnostics=TRUE)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + ssiratio,
## data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.7616 -0.7529 0.1275 0.8959 4.0723
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.787170 0.255523 26.562 < 2e-16 ***
## hi_empunion -0.897591 0.207991 -4.316 1.61e-05 ***
## totchr 0.450266 0.010422 43.201 < 2e-16 ***
## age -0.013218 0.002876 -4.596 4.36e-06 ***
## female -0.020406 0.031552 -0.647 0.518
## blhisp -0.217424 0.038688 -5.620 1.96e-08 ***
## linc 0.087002 0.022022 3.951 7.85e-05 ***
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 1 10082 183.98 < 2e-16 ***
## Wu-Hausman 1 10081 25.32 4.93e-07 ***
## Sargan 0 NA NA NA
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.318 on 10082 degrees of freedom
## Multiple R-Squared: 0.06395, Adjusted R-squared: 0.0634
## Wald test: 319.6 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(todas, type = 'text', title = "Prueba Hausman", df = F, digits = 4)##
## Prueba Hausman
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.8976***
## (0.2080)
##
## totchr 0.4503***
## (0.0104)
##
## age -0.0132***
## (0.0029)
##
## female -0.0204
## (0.0316)
##
## blhisp -0.2174***
## (0.0387)
##
## linc 0.0870***
## (0.0220)
##
## Constant 6.7872***
## (0.2555)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0640
## Adjusted R2 0.0634
## Residual Std. Error 1.3182
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01H0: El regresor es exógeno, es decir, es preferible usar MCO H1:El regresor es endógeno, es decir, es preferble usar MCO2E
Nuestro valor p, en Wu-hausman es de 4.93e-07, es decir, que rechazamos la nula y por tanto nos iremos por MCO2E.
comb1 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + ssiratio + lowincome, data=mus06data)
summary(comb1,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + ssiratio +
## lowincome, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.6790 -0.7315 0.1408 0.8792 3.9443
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.610127 0.240621 27.471 < 2e-16 ***
## hi_empunion -0.711863 0.190571 -3.735 0.000188 ***
## totchr 0.448376 0.010176 44.060 < 2e-16 ***
## age -0.011365 0.002723 -4.174 3.02e-05 ***
## female -0.005453 0.030342 -0.180 0.857372
## blhisp -0.204784 0.037556 -5.453 5.08e-08 ***
## linc 0.072372 0.020805 3.479 0.000506 ***
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 2 10081 105.261 < 2e-16 ***
## Wu-Hausman 1 10081 18.948 1.36e-05 ***
## Sargan 1 NA 6.721 0.00953 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.29 on 10082 degrees of freedom
## Multiple R-Squared: 0.103, Adjusted R-squared: 0.1025
## Wald test: 332.6 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb1, type = 'text', title = "Combinación de Instrumentos 1", df = F, digits = 4)##
## Combinación de Instrumentos 1
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.7119***
## (0.1906)
##
## totchr 0.4484***
## (0.0102)
##
## age -0.0114***
## (0.0027)
##
## female -0.0055
## (0.0303)
##
## blhisp -0.2048***
## (0.0376)
##
## linc 0.0724***
## (0.0208)
##
## Constant 6.6101***
## (0.2406)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.1030
## Adjusted R2 0.1025
## Residual Std. Error 1.2904
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01comb2 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + ssiratio + lowincome + firmsz, data=mus06data)
summary(comb2,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + ssiratio +
## lowincome + firmsz, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.7179 -0.7447 0.1324 0.8832 4.0046
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.693514 0.240547 27.826 < 2e-16 ***
## hi_empunion -0.799340 0.189070 -4.228 2.38e-05 ***
## totchr 0.449266 0.010268 43.753 < 2e-16 ***
## age -0.012238 0.002726 -4.489 7.23e-06 ***
## female -0.012496 0.030500 -0.410 0.682030
## blhisp -0.210737 0.037840 -5.569 2.63e-08 ***
## linc 0.079263 0.020817 3.808 0.000141 ***
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 3 10080 72.72 < 2e-16 ***
## Wu-Hausman 1 10081 24.27 8.52e-07 ***
## Sargan 2 NA 12.61 0.00183 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.303 on 10082 degrees of freedom
## Multiple R-Squared: 0.08566, Adjusted R-squared: 0.08511
## Wald test: 327 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb2, type = 'text', title = "Combinación de Instrumentos 2", df = F, digits = 4)##
## Combinación de Instrumentos 2
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.7993***
## (0.1891)
##
## totchr 0.4493***
## (0.0103)
##
## age -0.0122***
## (0.0027)
##
## female -0.0125
## (0.0305)
##
## blhisp -0.2107***
## (0.0378)
##
## linc 0.0793***
## (0.0208)
##
## Constant 6.6935***
## (0.2405)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0857
## Adjusted R2 0.0851
## Residual Std. Error 1.3028
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01comb3 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + ssiratio + lowincome + firmsz + multlc, data=mus06data)
summary(comb3,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + ssiratio +
## lowincome + firmsz + multlc, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.7459 -0.7523 0.1286 0.8935 4.0480
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.753569 0.233459 28.928 < 2e-16 ***
## hi_empunion -0.862342 0.177856 -4.849 1.26e-06 ***
## totchr 0.449907 0.010321 43.590 < 2e-16 ***
## age -0.012866 0.002661 -4.835 1.35e-06 ***
## female -0.017568 0.030232 -0.581 0.561
## blhisp -0.215025 0.037838 -5.683 1.36e-08 ***
## linc 0.084225 0.020272 4.155 3.28e-05 ***
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 4 10079 62.75 < 2e-16 ***
## Wu-Hausman 1 10081 32.12 1.49e-08 ***
## Sargan 3 NA 13.27 0.00408 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.312 on 10082 degrees of freedom
## Multiple R-Squared: 0.07201, Adjusted R-squared: 0.07145
## Wald test: 323.2 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb3, type = 'text', title = "Combinación de Instrumentos 3", df = F, digits = 4)##
## Combinación de Instrumentos 3
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.8623***
## (0.1779)
##
## totchr 0.4499***
## (0.0103)
##
## age -0.0129***
## (0.0027)
##
## female -0.0176
## (0.0302)
##
## blhisp -0.2150***
## (0.0378)
##
## linc 0.0842***
## (0.0203)
##
## Constant 6.7536***
## (0.2335)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0720
## Adjusted R2 0.0715
## Residual Std. Error 1.3125
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01comb4 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + ssiratio + firmsz, data=mus06data)
summary(comb4,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + ssiratio +
## firmsz, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.8025 -0.7658 0.1167 0.9097 4.1356
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.874762 0.255549 26.902 < 2e-16 ***
## hi_empunion -0.989480 0.206376 -4.795 1.65e-06 ***
## totchr 0.451201 0.010538 42.816 < 2e-16 ***
## age -0.014134 0.002881 -4.906 9.43e-07 ***
## female -0.027804 0.031753 -0.876 0.381
## blhisp -0.223678 0.039047 -5.728 1.04e-08 ***
## linc 0.094240 0.022045 4.275 1.93e-05 ***
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 2 10081 95.731 < 2e-16 ***
## Wu-Hausman 1 10081 31.599 1.95e-08 ***
## Sargan 1 NA 4.958 0.026 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.334 on 10082 degrees of freedom
## Multiple R-Squared: 0.04156, Adjusted R-squared: 0.04099
## Wald test: 313 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb4, type = 'text', title = "Combinación de Instrumentos 4", df = F, digits = 4)##
## Combinación de Instrumentos 4
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.9895***
## (0.2064)
##
## totchr 0.4512***
## (0.0105)
##
## age -0.0141***
## (0.0029)
##
## female -0.0278
## (0.0318)
##
## blhisp -0.2237***
## (0.0390)
##
## linc 0.0942***
## (0.0220)
##
## Constant 6.8748***
## (0.2555)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0416
## Adjusted R2 0.0410
## Residual Std. Error 1.3338
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01comb5 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + ssiratio + lowincome + multlc, data=mus06data)
summary(comb5,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + ssiratio +
## lowincome + multlc, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.7278 -0.7456 0.1300 0.8871 4.0200
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.714783 0.233028 28.815 < 2e-16 ***
## hi_empunion -0.821653 0.178000 -4.616 3.96e-06 ***
## totchr 0.449493 0.010273 43.754 < 2e-16 ***
## age -0.012460 0.002655 -4.694 2.72e-06 ***
## female -0.014292 0.030125 -0.474 0.635
## blhisp -0.212256 0.037678 -5.633 1.81e-08 ***
## linc 0.081020 0.020229 4.005 6.24e-05 ***
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 3 10080 82.712 < 2e-16 ***
## Wu-Hausman 1 10081 29.042 7.24e-08 ***
## Sargan 2 NA 8.741 0.0126 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.306 on 10082 degrees of freedom
## Multiple R-Squared: 0.08093, Adjusted R-squared: 0.08039
## Wald test: 325.9 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb5, type = 'text', title = "Combinación de Instrumentos 5", df = F, digits = 4)##
## Combinación de Instrumentos 5
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.8217***
## (0.1780)
##
## totchr 0.4495***
## (0.0103)
##
## age -0.0125***
## (0.0027)
##
## female -0.0143
## (0.0301)
##
## blhisp -0.2123***
## (0.0377)
##
## linc 0.0810***
## (0.0202)
##
## Constant 6.7148***
## (0.2330)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0809
## Adjusted R2 0.0804
## Residual Std. Error 1.3062
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01comb6 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + ssiratio + multlc, data=mus06data)
summary(comb6,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + ssiratio +
## multlc, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.8027 -0.7657 0.1166 0.9099 4.1359
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.875188 0.245362 28.021 < 2e-16 ***
## hi_empunion -0.989927 0.192280 -5.148 2.68e-07 ***
## totchr 0.451205 0.010511 42.927 < 2e-16 ***
## age -0.014138 0.002782 -5.082 3.81e-07 ***
## female -0.027840 0.031176 -0.893 0.372
## blhisp -0.223709 0.038714 -5.778 7.76e-09 ***
## linc 0.094275 0.021241 4.438 9.16e-06 ***
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 2 10081 110.613 < 2e-16 ***
## Wu-Hausman 1 10081 36.561 1.53e-09 ***
## Sargan 1 NA 1.164 0.281
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.334 on 10082 degrees of freedom
## Multiple R-Squared: 0.04145, Adjusted R-squared: 0.04088
## Wald test: 313.5 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb6, type = 'text', title = "Combinación de Instrumentos 6", df = F, digits = 4)##
## Combinación de Instrumentos 6
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.9899***
## (0.1923)
##
## totchr 0.4512***
## (0.0105)
##
## age -0.0141***
## (0.0028)
##
## female -0.0278
## (0.0312)
##
## blhisp -0.2237***
## (0.0387)
##
## linc 0.0943***
## (0.0212)
##
## Constant 6.8752***
## (0.2454)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0414
## Adjusted R2 0.0409
## Residual Std. Error 1.3339
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01comb7 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + ssiratio + lowincome + multlc, data=mus06data)
summary(comb7,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + ssiratio +
## lowincome + multlc, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.7278 -0.7456 0.1300 0.8871 4.0200
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.714783 0.233028 28.815 < 2e-16 ***
## hi_empunion -0.821653 0.178000 -4.616 3.96e-06 ***
## totchr 0.449493 0.010273 43.754 < 2e-16 ***
## age -0.012460 0.002655 -4.694 2.72e-06 ***
## female -0.014292 0.030125 -0.474 0.635
## blhisp -0.212256 0.037678 -5.633 1.81e-08 ***
## linc 0.081020 0.020229 4.005 6.24e-05 ***
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 3 10080 82.712 < 2e-16 ***
## Wu-Hausman 1 10081 29.042 7.24e-08 ***
## Sargan 2 NA 8.741 0.0126 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.306 on 10082 degrees of freedom
## Multiple R-Squared: 0.08093, Adjusted R-squared: 0.08039
## Wald test: 325.9 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb7, type = 'text', title = "Combinación de Instrumentos 7", df = F, digits = 4)##
## Combinación de Instrumentos 7
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.8217***
## (0.1780)
##
## totchr 0.4495***
## (0.0103)
##
## age -0.0125***
## (0.0027)
##
## female -0.0143
## (0.0301)
##
## blhisp -0.2123***
## (0.0377)
##
## linc 0.0810***
## (0.0202)
##
## Constant 6.7148***
## (0.2330)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.0809
## Adjusted R2 0.0804
## Residual Std. Error 1.3062
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01comb8 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + multlc + lowincome, data=mus06data)
summary(comb8,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + multlc +
## lowincome, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.6382 -0.7230 0.1409 0.8645 3.8811
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.522635 0.294650 22.137 < 2e-16 ***
## hi_empunion -0.620079 0.261996 -2.367 0.01796 *
## totchr 0.447442 0.010252 43.646 < 2e-16 ***
## age -0.010450 0.003250 -3.216 0.00131 **
## female 0.001936 0.033433 0.058 0.95382
## blhisp -0.198537 0.039213 -5.063 4.2e-07 ***
## linc 0.065143 0.025093 2.596 0.00944 **
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 2 10081 54.139 < 2e-16 ***
## Wu-Hausman 1 10081 7.593 0.00587 **
## Sargan 1 NA 8.059 0.00453 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.279 on 10082 degrees of freedom
## Multiple R-Squared: 0.1193, Adjusted R-squared: 0.1188
## Wald test: 337.3 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb8, type = 'text', title = "Combinación de Instrumentos 8", df = F, digits = 4)##
## Combinación de Instrumentos 8
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.6201**
## (0.2620)
##
## totchr 0.4474***
## (0.0103)
##
## age -0.0105***
## (0.0032)
##
## female 0.0019
## (0.0334)
##
## blhisp -0.1985***
## (0.0392)
##
## linc 0.0651***
## (0.0251)
##
## Constant 6.5226***
## (0.2947)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.1193
## Adjusted R2 0.1188
## Residual Std. Error 1.2786
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01comb9 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + multlc + firmsz, data=mus06data)
summary(comb9,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + multlc +
## firmsz, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -7.03222 -0.86487 0.08313 0.98693 4.54685
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 7.367225 0.421170 17.492 < 2e-16 ***
## hi_empunion -1.506102 0.401173 -3.754 0.000175 ***
## totchr 0.456457 0.011895 38.373 < 2e-16 ***
## age -0.019286 0.004557 -4.233 2.33e-05 ***
## female -0.069396 0.043593 -1.592 0.111432
## blhisp -0.258839 0.047952 -5.398 6.89e-08 ***
## linc 0.134933 0.035470 3.804 0.000143 ***
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 2 10081 29.273 2.11e-13 ***
## Wu-Hausman 1 10081 21.311 3.95e-06 ***
## Sargan 1 NA 2.607 0.106
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.443 on 10082 degrees of freedom
## Multiple R-Squared: -0.122, Adjusted R-squared: -0.1226
## Wald test: 266.4 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb9, type = 'text', title = "Combinación de Instrumentos 9", df = F, digits = 4)##
## Combinación de Instrumentos 9
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -1.5061***
## (0.4012)
##
## totchr 0.4565***
## (0.0119)
##
## age -0.0193***
## (0.0046)
##
## female -0.0694
## (0.0436)
##
## blhisp -0.2588***
## (0.0480)
##
## linc 0.1349***
## (0.0355)
##
## Constant 7.3672***
## (0.4212)
##
## -----------------------------------------------
## Observations 10,089
## R2 -0.1220
## Adjusted R2 -0.1226
## Residual Std. Error 1.4432
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01comb10 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + firmsz + lowincome, data=mus06data)
summary(comb10,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + firmsz +
## lowincome, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.5206 -0.6911 0.1484 0.8526 3.6989
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.270672 0.351869 17.821 < 2e-16 ***
## hi_empunion -0.355754 0.332325 -1.071 0.2844
## totchr 0.444752 0.010270 43.307 < 2e-16 ***
## age -0.007814 0.003817 -2.047 0.0406 *
## female 0.023216 0.036898 0.629 0.5292
## blhisp -0.180547 0.041012 -4.402 1.08e-05 ***
## linc 0.044323 0.029677 1.493 0.1353
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 2 10081 32.151 1.21e-14 ***
## Wu-Hausman 1 10081 1.727 0.188783
## Sargan 1 NA 11.094 0.000866 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.253 on 10082 degrees of freedom
## Multiple R-Squared: 0.1549, Adjusted R-squared: 0.1544
## Wald test: 350.8 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb10, type = 'text', title = "Combinación de Instrumentos 10", df = F, digits = 4)##
## Combinación de Instrumentos 10
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.3558
## (0.3323)
##
## totchr 0.4448***
## (0.0103)
##
## age -0.0078**
## (0.0038)
##
## female 0.0232
## (0.0369)
##
## blhisp -0.1805***
## (0.0410)
##
## linc 0.0443
## (0.0297)
##
## Constant 6.2707***
## (0.3519)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.1549
## Adjusted R2 0.1544
## Residual Std. Error 1.2525
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01comb11 <- ivreg(ldrugexp ~ hi_empunion + totchr + age + female + blhisp + linc | totchr + age + female + blhisp + linc + multlc + firmsz+ lowincome, data=mus06data)
summary(comb11,diagnostics = T)##
## Call:
## ivreg(formula = ldrugexp ~ hi_empunion + totchr + age + female +
## blhisp + linc | totchr + age + female + blhisp + linc + multlc +
## firmsz + lowincome, data = mus06data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.6857 -0.7327 0.1375 0.8805 3.9547
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.624424 0.294534 22.491 < 2e-16 ***
## hi_empunion -0.726861 0.260735 -2.788 0.005318 **
## totchr 0.448528 0.010352 43.329 < 2e-16 ***
## age -0.011515 0.003252 -3.541 0.000401 ***
## female -0.006661 0.033587 -0.198 0.842802
## blhisp -0.205805 0.039511 -5.209 1.94e-07 ***
## linc 0.073554 0.025099 2.931 0.003392 **
##
## Diagnostic tests:
## df1 df2 statistic p-value
## Weak instruments 3 10080 37.24 < 2e-16 ***
## Wu-Hausman 1 10081 10.44 0.00124 **
## Sargan 2 NA 13.20 0.00136 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.292 on 10082 degrees of freedom
## Multiple R-Squared: 0.1002, Adjusted R-squared: 0.09965
## Wald test: 330.6 on 6 and 10082 DF, p-value: < 2.2e-16stargazer(comb11, type = 'text', title = "Combinación de Instrumentos 11", df = F, digits = 4)##
## Combinación de Instrumentos 11
## ===============================================
## Dependent variable:
## ---------------------------
## ldrugexp
## -----------------------------------------------
## hi_empunion -0.7269***
## (0.2607)
##
## totchr 0.4485***
## (0.0104)
##
## age -0.0115***
## (0.0033)
##
## female -0.0067
## (0.0336)
##
## blhisp -0.2058***
## (0.0395)
##
## linc 0.0736***
## (0.0251)
##
## Constant 6.6244***
## (0.2945)
##
## -----------------------------------------------
## Observations 10,089
## R2 0.1002
## Adjusted R2 0.0997
## Residual Std. Error 1.2924
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01Ahora que ya realizamos todas las combinaciones posibles elegimos la número 6. Primero porque el valor p es mayor a 0.05 y de acuerdo a la regla de decisión de la Prueba de Sargan:
H~0: sobreidentificación válida H~1: sobreidentificación no válida
p > 0.05 no se rechaza la hipótesis nulaOtra razón por la que elegimos esa regresión fue porque inicialmente el instrumento “ssratio” era muy fuerte.
Interpretación:
El desembolso total en medicamentos recetados de una persona que tenga seguro de salud patrocinado por el empleador o por el sindicato es menor en 98.99% que una persona que no tiene seguro de salud.
Por cada enfermedad crónica que se padezca, el desembolso total en medicamentos recetados aumenta un 45%
Por cada año más que tenga la persona, el desembolso total en medicamentos recetados disminuye en 1.41%
El desembolso total en medicamentos recetados de una mujer es 2.78% menor que el de un hombre y, a diferencia de las demás variables, esta es una variable no significativa
El desembolso total en medicamentos recetados de una persona negra o hispana es 22.37% menor que una persona que no lo es
Por cada unidad porcentual que aumente el ingreso anual del hogar, el desembolso total en medicamentos recetados aumenta 9.4%,
Cuando todas las variables dependientes valen 0, el desembolso total en medicamentos recetados es de 687.51% y es una variable significativa.
Cabe resaltar que cada interpretación es en promedio, ceteris paribus, es decir, que por cada variación porcentual que estemos presentando, todo lo demás permanecerá constante.