Hi, now for something completely different: As I mentioned in passing at the conference, I have received some (mild) complaints about the output of the ADF test in gretl. And indeed I think the arrangement of the information is suboptimal. So here's a concrete suggestion with before-and-after [need to view this with monospaced font]: Now / before: " Augmented Dickey-Fuller test for LRM including 4 lags of (1-L)LRM sample size 50 unit-root null hypothesis: a = 1 test with constant model: (1-L)y = b0 + (a-1)*y(-1) + ... + e 1st-order autocorrelation coeff. for e: 0.075 lagged differences: F(4, 44) = 4.290 [0.0051] estimated value of (a - 1): -0.0591989 test statistic: tau_c(1) = -1.70189 asymptotic p-value 0.4303 " And my suggestion / after: " LRM: Augmented Dickey-Fuller test (H0: unit root, a = 1) ------------------------------------- Test with constant â - 1 = -0.0591989 Test stat tau_c(1) = -1.70189 Asympt. p-value = 0.4303 ------------------------------------- Test equation: Dependent variable differenced: (1-L)LRM, lagged level with coefficient a - 1, including 4 lagged differences (exclusion test F(4, 44) = 4.290 [0.0051]), and constant term. Residual 1st-order autocorr = 0.075 Sample size = 50 -------------------------------------- " Now I'm not saying my suggestion is optimal, but I do think it would be an improvement. If somebody directs me to the right places in the source, I could try to make the corresponding changes. What do you think?

But how about like this (but there's a problem, see below): " ... Test equation: (1-L)LRM = b0 + (a-1)*LRM(-1) + ... + e Dependent variable differenced: (1-L)LRM, constant term, lagged level with coefficient a - 1, including 4 lagged differences (exclusion test F(4, 44) = 4.290 [0.0051]). Residual 1st-order autocorr = 0.075 Sample size = 50 -------------------------------------- " Note that I have substituted the variable name "LRM" for the generic "y". I think this is much better because "y" is another undefined symbol and has to be understood from this specific context which IMO adds a layer of confusion. (Think about if you have another variable named "y" in your dataset, but you're really testing x...) However, this name substitution creates problems if the variable name is long. Right now I don't know what to do about it... ? thanks, sven

For now I've committed a minimal change to ADF output (without prejudice to more radical change later): I've moved the supplementary info (rho-hat for the residual and the joint significance of lagged differences, if any) below the main output, namely the estimate of (a - 1), the tau statistic and its p-value.

On the "(1-L)" notation in the model string, to which Sven raised an objection: I'd be kinda surprised if someone executing a unit root test didn't know 'L' as the lag operator in this context.

Am 29.06.2015 um 14:03 schrieb Sven Schreiber: > Am 23.06.2015 um 17:06 schrieb Allin Cottrell: > > >> >> For now I've committed a minimal change to ADF output (without prejudice >> to more radical change later): I've moved the supplementary info >> (rho-hat for the residual and the joint significance of lagged >> differences, if any) below the main output, namely the estimate of (a - >> 1), the tau statistic and its p-value. > > Thanks Allin, I still have the more radical formatting on my to-do-list. > I guess it'll take some time to become acquainted with the way printout > is set up in the source, so I cannot do this right away, but will try. > Ok here's a diff file for adf_kpss.c -- I spent a ridiculous amount of time ending up not too far from the previous status quo, but well. Guess I became acquainted a little with the C source style along the way. (If again the encoding is messed up I have no idea why. I re-checked and every tool seems to be set to utf-8 here, including thunderbird AFAICS.) I'm also attaching the actual output of how it looks with this change. Better double-check the changes, because while I did test, I also managed to produce a few crashes through typos and syntax mistakes...