Heteroscedasticity

[-wish-]

Hello,
I have 72 monthly observations, I took days to foud a model of ARMA
Finally I founded this equation " dliq c @trend ar(2) ar(4) ar(5) ar(10) ar(12) ma(4) ma(5) ma(10) ma(12) " I had this correlogramm:
Included observations: 59
Q-statistic probabilities adjusted for 9 ARMA term(s)

Autocorrelation Partial Correlation AC PAC Q-Stat Prob

. | . | . | . | 1 -0.016 -0.016 0.0163
. | . | . | . | 2 0.016 0.016 0.0324
. | . | . | . | 3 0.049 0.050 0.1876
. | . | . | . | 4 -0.009 -0.008 0.1928
. | . | . | . | 5 -0.035 -0.037 0.2736
. | . | . | . | 6 -0.026 -0.030 0.3207
. | . | . | . | 7 0.018 0.019 0.3424
. | . | . | . | 8 0.002 0.007 0.3427
. | . | . | . | 9 -0.026 -0.025 0.3917
. | . | . | . | 10 -0.015 -0.020 0.4084 0.523
. | . | . | . | 11 -0.046 -0.048 0.5643 0.754
. | . | . | . | 12 -0.005 -0.002 0.5659 0.904
. | . | . | . | 13 0.018 0.022 0.5901 0.964
. | . | . | . | 14 -0.038 -0.035 0.7070 0.983
. | . | . | . | 15 -0.010 -0.016 0.7154 0.994

But according to "ARCH" model, probabilities are very high!!
Heteroskedasticity Test: ARCH

F-statistic 0.130452 Prob. F(1,56) 0.7193
Obs*R-squared 0.134797 Prob. Chi-Square(1) 0.7135
Test Equation:
Dependent Variable: RESID^2
Method: Least Squares
Date: 03/04/13 Time: 18:01
Sample (adjusted): 2008M03 2012M12
Included observations: 58 after adjustments

Variable Coefficient Std. Error t-Statistic Prob.

C 13.58422 3.046560 4.458872 0.0000
RESID^2(-1) -0.051147 0.141609 -0.361182 0.7193

R-squared 0.002324 Mean dependent var 12.97538
Adjusted R-squared -0.015492 S.D. dependent var 19.17866
S.E. of regression 19.32664 Akaike info criterion 8.794721
Sum squared resid 20917.07 Schwarz criterion 8.865770
Log likelihood -253.0469 Hannan-Quinn criter. 8.822396
F-statistic 0.130452 Durbin-Watson stat 1.883763
Prob(F-statistic) 0.719323


Should I use another test ?? How can I correct heteroscedasticity ?
Thank you