Sample size

[Jone]

Hi there

I'm running a VAR model with five variables for the period 2000-2011. That's to say, 12 observations. I know the number of observations is small but the purpose of the research is to examine only this period. I succeeded in running all residual diagnosis tests except heteroscedasticity. I will be grateful if someone provide me with a name of a reference that support using small sample. ALso, any views and advice are welcomed. Looking forward to hearing from you soon.