1
$\begingroup$

all, I am new to this forum. I have a question of fixed effects in R....So I am trying to use plm function to find fixed effects like following:

> plm(PM25~policy+1,data=subset(part2,Delhi==1),model="within"
> ,index=c("station_id","date"))%>%   summary()

The results I get:

> Oneway (individual) effect Within Model
> 
> Call: plm(formula = PM25 ~ policy + 1, data = subset(part2, Delhi == 
>     1), model = "within", index = c("station_id", "date"))
> 
> Unbalanced Panel: n = 7, T = 73-159, N = 992
> 
> Residuals:
>      Min.   1st Qu.    Median   3rd Qu.      Max. 
> -193.7049  -54.6776   -9.6843   54.7431  318.4094 
> 
> Coefficients:
>        Estimate Std. Error t-value  Pr(>|t|)     policy  76.8167     9.9787  7.6981 3.354e-14 ***
> --- Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
> 
> Total Sum of Squares:    6846900 Residual Sum of Squares: 6458000
> R-Squared:      0.056803 Adj. R-Squared: 0.050093 F-statistic: 59.2603
> on 1 and 984 DF, p-value: 3.3535e-14

I am wondering how can I find the intercept?

$\endgroup$
1
$\begingroup$

The within model you specify amounts to fitting a separate intercept for each unit in your panel data set, as in $$ y_{it}=a_i+\beta x_{it}+u_{it} $$ See, e.g., these posts for some context:

Difference between fixed effects dummies and fixed effects estimator?

How exactly does a "random effects model" in econometrics relate to mixed models outside of econometrics?

Hence, you cannot additionally fit an overall intercept $a$, as you would then have perfect collinearity between the $a_i$ and $a$, as the sum of the $a_i$ is identically one.

You can retrieve these estimates as follows:

data("Grunfeld", package = "plm")

wi <- plm(inv ~ value + capital,
          data = Grunfeld, model = "within", effect = "twoways")

fixef(wi)

One might of course leave out one of the intercepts. Something related to what you seek seems to be achieved by this command: https://rdrr.io/rforge/plm/man/within_intercept.html

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.