xtserial and xttest1 results are contradictory

[Kim JongBin]

hello world .

I have a question and I`m trying to solve this contradictory but I can`t

here`s my question

when i use xttest1 the result is...

    qui xtreg y x1 x2, re

    xttest1


     Estimated results:
                         |       Var     sd = sqrt(Var)
                ---------+-----------------------------
                       y  |   .2806127       .5297289
                       e |   .2234992      .47275704
                       u |   .0570193      .23878708

        Tests:
           Random Effects, Two Sided:
           ALM(Var(u)=0)         =   27.69 Pr>chi2(1) =  0.0000

           Random Effects, One Sided:
           ALM(Var(u)=0)         =    5.26 Pr>N(0,1)  =  0.0000

           Serial Correlation:
           ALM(lambda=0)         =   14.96 Pr>chi2(1) =  0.0001

           Joint Test:
           LM(Var(u)=0,lambda=0) =  144.52 Pr>chi2(2) =  0.0000

And then when I used xtserial the result is...

qui xtreg y x1 x2, re  

xtserial  y x1 x2


Wooldridge test for autocorrelation in panel data
H0: no first-order autocorrelation
    F(  1,     522) =      0.400

           Prob > F =      0.5273

Well, All I know is that If -xttest1-`s result is significant it has first-order autocorrelation. and if -xtserial-`s result isn`t significant it has first-order autocorrelation.

The result is contradictory. but stata is not wrong, so maybe i was wrong....

but i can`t understand this result.

my code was wrong..? or my Statistical memory is wrong...?