Dear R users,

I am applying the augmented-Dickey-Fuller Unit Root Test

(ur.df function of the urca package) to a time series of

approximately 50 values.

To be sure I understood what was going on with the ur.df

function, I checked the critical values of the 3 test

statistics (tau, phi2 and phi3 if a trend is included) or

the 2 test statistics (tau and phi1 if only a drift is

included) with the values found in Hamilton (1994) and

Dickey and Fuller (1981).

The ur.df function critical values correspond to Hamilton's

critical values for a sample of size 100. I was surprised

that it does not correspond to the values for a sample of

size 50.

I was then wondering if there was an error somewhere or if I

did not understand something. Since I am not an expert in

that field, I may have misunderstood something.

Thank you very much for your help!

With kind regards,

S?verine Gaille

_________________________

Ecole des HEC

Universite de Lausanne

CH-1015 Lausanne

Tel (+41 21) 692 33 72