##İLERİ PANEL VERİ YÖNTEMLERİ ###SABİT ETKİLER TAHMİNİ
library(rmarkdown)
library(wooldridge)
library(plm)Burada panel verileri analiz etmek için çoklu regresyonlar için kullandığımız lm() komutu yerine plm paketini kullanacağız. Öncelikle plm paketini R-studio’ya yüklüyoruz.
data("jtrain")Verisetini gözlemleyebilmek ve değişkenlerimin tanımlarını anlayabilmek için data komutuyla verisetini indiriyorum.
head(jtrain)## year fcode employ sales avgsal scrap rework tothrs union grant d89 d88
## 1 1987 410032 100 47000000 35000 NA NA 12 0 0 0 0
## 2 1988 410032 131 43000000 37000 NA NA 8 0 0 0 1
## 3 1989 410032 123 49000000 39000 NA NA 8 0 0 1 0
## 4 1987 410440 12 1560000 10500 NA NA 12 0 0 0 0
## 5 1988 410440 13 1970000 11000 NA NA 12 0 0 0 1
## 6 1989 410440 14 2350000 11500 NA NA 10 0 0 1 0
## totrain hrsemp lscrap lemploy lsales lrework lhrsemp lscrap_1 grant_1
## 1 100 12.000000 NA 4.605170 17.66566 NA 2.564949 NA 0
## 2 50 3.053435 NA 4.875197 17.57671 NA 1.399565 NA 0
## 3 50 3.252033 NA 4.812184 17.70733 NA 1.447397 NA 0
## 4 12 12.000000 NA 2.484907 14.26020 NA 2.564949 NA 0
## 5 13 12.000000 NA 2.564949 14.49354 NA 2.564949 NA 0
## 6 14 10.000000 NA 2.639057 14.66993 NA 2.397895 NA 0
## clscrap cgrant clemploy clsales lavgsal clavgsal cgrant_1
## 1 NA 0 NA NA 10.463103 NA NA
## 2 NA 0 0.27002716 -0.0889492 10.518673 0.05556965 0
## 3 NA 0 -0.06301308 0.1306210 10.571317 0.05264378 0
## 4 NA 0 NA NA 9.259130 NA NA
## 5 NA 0 0.08004260 0.2333469 9.305651 0.04652023 0
## 6 NA 0 0.07410812 0.1763821 9.350102 0.04445171 0
## chrsemp clhrsemp
## 1 NA NA
## 2 -8.9465647 -1.16538453
## 3 0.1985974 0.04783237
## 4 NA NA
## 5 0.0000000 0.00000000
## 6 -2.0000000 -0.16705394tail(jtrain)## year fcode employ sales avgsal scrap rework tothrs union grant d89 d88
## 466 1987 419483 133 11000000 13957 20 NA 0 1 0 0 0
## 467 1988 419483 108 11500000 14810 25 NA 0 1 0 0 1
## 468 1989 419483 129 12000000 14227 30 NA 20 1 0 1 0
## 469 1987 419486 80 7000000 16000 NA NA 0 0 0 0 0
## 470 1988 419486 90 8500000 17000 NA NA 0 0 0 0 1
## 471 1989 419486 100 9900000 18000 NA NA 40 0 1 1 0
## totrain hrsemp lscrap lemploy lsales lrework lhrsemp lscrap_1
## 466 0 0.000000 2.995732 4.890349 16.21341 NA 0.000000 NA
## 467 0 0.000000 3.218876 4.682131 16.25786 NA 0.000000 2.995732
## 468 20 3.100775 3.401197 4.859812 16.30042 NA 1.411176 3.218876
## 469 0 0.000000 NA 4.382027 15.76142 NA 0.000000 NA
## 470 0 0.000000 NA 4.499810 15.95558 NA 0.000000 NA
## 471 90 36.000000 NA 4.605170 16.10805 NA 3.610918 NA
## grant_1 clscrap cgrant clemploy clsales lavgsal clavgsal
## 466 0 NA 0 NA NA 9.543736 NA
## 467 0 0.2231436 0 -0.2082176 0.04445267 9.603058 0.05932140
## 468 0 0.1823215 0 0.1776810 0.04255867 9.562897 -0.04016113
## 469 0 NA 0 NA NA 9.680344 NA
## 470 0 NA 0 0.1177831 0.19415665 9.740969 0.06062508
## 471 0 NA 1 0.1053605 0.15246868 9.798127 0.05715847
## cgrant_1 chrsemp clhrsemp
## 466 NA NA NA
## 467 0 0.000000 0.000000
## 468 0 3.100775 1.411176
## 469 NA NA NA
## 470 0 0.000000 0.000000
## 471 0 36.000000 3.610918“Head” komutu ile ilk 6 gözlemi “tail” komutuyla ise son 6 gözlemi tablo şekilinde gösterdik.
###Veri Setini Panel Veri Setine Cevirme
indexim <- pdata.frame(jtrain , index = c( "fcode","year" ))pdim(indexim)## Balanced Panel: n = 157, T = 3, N = 471plm paketinin içinde bulunan pdim komutu sayesinde verisetinin balansını kontrol edebilir, kaç kişi için toplam kaç yıl veri toplandığını görebiliriz. n burada 157 kişiden, T, 3 yıl boyunca toplam 471 tane gözlem toplanıldığını göstermektedir.
summary(jtrain)## year fcode employ sales
## Min. :1987 Min. :410032 Min. : 4.00 Min. : 110000
## 1st Qu.:1987 1st Qu.:410604 1st Qu.: 15.00 1st Qu.: 1550000
## Median :1988 Median :418084 Median : 30.00 Median : 3000000
## Mean :1988 Mean :415709 Mean : 59.32 Mean : 6116037
## 3rd Qu.:1989 3rd Qu.:419309 3rd Qu.: 72.00 3rd Qu.: 7700000
## Max. :1989 Max. :419486 Max. :525.00 Max. :54000000
## NA's :31 NA's :98
## avgsal scrap rework tothrs
## Min. : 4237 Min. : 0.0100 Min. : 0.000 Min. : 0.0
## 1st Qu.:14102 1st Qu.: 0.5925 1st Qu.: 0.350 1st Qu.: 0.0
## Median :17773 Median : 1.4150 Median : 1.160 Median : 12.0
## Mean :18873 Mean : 3.8436 Mean : 3.474 Mean : 29.2
## 3rd Qu.:22360 3rd Qu.: 4.0000 3rd Qu.: 4.000 3rd Qu.: 40.0
## Max. :42583 Max. :30.0000 Max. :40.000 Max. :320.0
## NA's :65 NA's :309 NA's :348 NA's :56
## union grant d89 d88
## Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000
## 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000
## Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000
## Mean :0.1975 Mean :0.1401 Mean :0.3333 Mean :0.3333
## 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:1.0000
## Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000
##
## totrain hrsemp lscrap lemploy
## Min. : 0.00 Min. : 0.000 Min. :-4.6052 Min. :1.386
## 1st Qu.: 0.00 1st Qu.: 0.000 1st Qu.:-0.5234 1st Qu.:2.708
## Median : 8.00 Median : 3.308 Median : 0.3471 Median :3.401
## Mean : 23.09 Mean : 14.968 Mean : 0.3937 Mean :3.531
## 3rd Qu.: 25.00 3rd Qu.: 18.663 3rd Qu.: 1.3863 3rd Qu.:4.277
## Max. :350.00 Max. :163.917 Max. : 3.4012 Max. :6.263
## NA's :6 NA's :81 NA's :309 NA's :31
## lsales lrework lhrsemp lscrap_1
## Min. :11.61 Min. :-4.6052 Min. :0.000 Min. :-4.6052
## 1st Qu.:14.25 1st Qu.:-0.9163 1st Qu.:0.000 1st Qu.:-0.2675
## Median :14.91 Median : 0.1823 Median :1.460 Median : 0.4414
## Mean :15.03 Mean : 0.1642 Mean :1.650 Mean : 0.5129
## 3rd Qu.:15.86 3rd Qu.: 1.3863 3rd Qu.:2.979 3rd Qu.: 1.6094
## Max. :17.80 Max. : 3.6889 Max. :5.105 Max. : 3.4012
## NA's :98 NA's :350 NA's :81 NA's :363
## grant_1 clscrap cgrant clemploy
## Min. :0.00000 Min. :-3.3142 Min. :-1.00000 Min. :-0.98083
## 1st Qu.:0.00000 1st Qu.:-0.3975 1st Qu.: 0.00000 1st Qu.:-0.02899
## Median :0.00000 Median :-0.1411 Median : 0.00000 Median : 0.07066
## Mean :0.07643 Mean :-0.2211 Mean : 0.06369 Mean : 0.08202
## 3rd Qu.:0.00000 3rd Qu.: 0.0093 3rd Qu.: 0.00000 3rd Qu.: 0.18232
## Max. :1.00000 Max. : 2.3979 Max. : 1.00000 Max. : 1.67398
## NA's :363 NA's :181
## clsales lavgsal clavgsal cgrant_1
## Min. :-1.98287 Min. : 8.352 Min. :-0.40547 Min. :0.0000
## 1st Qu.:-0.01101 1st Qu.: 9.554 1st Qu.: 0.02228 1st Qu.:0.0000
## Median : 0.10711 Median : 9.785 Median : 0.05716 Median :0.0000
## Mean : 0.11587 Mean : 9.785 Mean : 0.06026 Mean :0.1147
## 3rd Qu.: 0.22314 3rd Qu.:10.015 3rd Qu.: 0.09076 3rd Qu.:0.0000
## Max. : 2.89670 Max. :10.659 Max. : 0.56891 Max. :1.0000
## NA's :226 NA's :65 NA's :204 NA's :157
## chrsemp clhrsemp
## Min. :-88.62255 Min. :-4.02535
## 1st Qu.: -0.07257 1st Qu.:-0.01493
## Median : 0.19860 Median : 0.03479
## Mean : 5.93591 Mean : 0.50370
## 3rd Qu.: 11.00952 3rd Qu.: 1.36811
## Max. :142.00000 Max. : 4.39445
## NA's :220 NA's :220PLM regresyon oluşturma
modelim <- plm(d89 ~ union + grant + d88 , data = indexim , model = "within" ) summary(modelim)## Oneway (individual) effect Within Model
##
## Call:
## plm(formula = d89 ~ union + grant + d88, data = indexim, model = "within")
##
## Balanced Panel: n = 157, T = 3, N = 471
##
## Residuals:
## Min. 1st Qu. Median 3rd Qu. Max.
## -0.515875 -0.397188 0.031751 0.484125 0.602812
##
## Coefficients:
## Estimate Std. Error t-value Pr(>|t|)
## grant 0.356063 0.074606 4.7726 2.799e-06 ***
## d88 -0.547626 0.048372 -11.3211 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Total Sum of Squares: 104.67
## Residual Sum of Squares: 73.159
## R-Squared: 0.30103
## Adj. R-Squared: -0.052938
## F-statistic: 67.1849 on 2 and 312 DF, p-value: < 2.22e-16LM REGRESYONU
Kesenli regreyon
modelim2 <- lm (d89 ~ union + grant + d88 , data = indexim)Regresyon özeti
summary(modelim2)##
## Call:
## lm(formula = d89 ~ union + grant + d88, data = indexim)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.47598 -0.47598 0.06066 0.52402 0.53491
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.47598 0.02476 19.226 < 2e-16 ***
## union -0.01089 0.04625 -0.236 0.814
## grant 0.27392 0.05393 5.079 5.5e-07 ***
## d88 -0.53664 0.03967 -13.526 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.3991 on 467 degrees of freedom
## Multiple R-squared: 0.2893, Adjusted R-squared: 0.2847
## F-statistic: 63.35 on 3 and 467 DF, p-value: < 2.2e-16Kesensiz Regresyon
modelim3 <- lm(d89 ~ union + grant + d88 -1 , data = indexim)Kesensiz regresyon özeti
summary(modelim3)##
## Call:
## lm(formula = d89 ~ union + grant + d88 - 1, data = indexim)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.6114 0.0000 0.1691 0.6915 1.0000
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## union 0.30851 0.05771 5.346 1.41e-07 ***
## grant 0.47202 0.07079 6.668 7.31e-11 ***
## d88 -0.16915 0.04648 -3.639 0.000304 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5336 on 468 degrees of freedom
## Multiple R-squared: 0.1511, Adjusted R-squared: 0.1457
## F-statistic: 27.77 on 3 and 468 DF, p-value: < 2.2e-16library(stargazer)##
## Please cite as:## Hlavac, Marek (2022). stargazer: Well-Formatted Regression and Summary Statistics Tables.## R package version 5.2.3. https://CRAN.R-project.org/package=stargazerstargazer(list(modelim, modelim2, modelim3), type = "text")##
## ===========================================================================================
## Dependent variable:
## -----------------------------------------------------------------------
## d89
## panel OLS
## linear
## (1) (2) (3)
## -------------------------------------------------------------------------------------------
## union -0.011 0.309***
## (0.046) (0.058)
##
## grant 0.356*** 0.274*** 0.472***
## (0.075) (0.054) (0.071)
##
## d88 -0.548*** -0.537*** -0.169***
## (0.048) (0.040) (0.046)
##
## Constant 0.476***
## (0.025)
##
## -------------------------------------------------------------------------------------------
## Observations 471 471 471
## R2 0.301 0.289 0.151
## Adjusted R2 -0.053 0.285 0.146
## Residual Std. Error 0.399 (df = 467) 0.534 (df = 468)
## F Statistic 67.185*** (df = 2; 312) 63.353*** (df = 3; 467) 27.769*** (df = 3; 468)
## ===========================================================================================
## Note: *p<0.1; **p<0.05; ***p<0.01