All Seasons wrote: ↑Tue Aug 25, 2020 5:51 pm
There are qualitative considerations and some people just refuse to believe it.
I have mentioned multiple times that there are other considerations, it's the data approach that doesn't work. If your argument is that gold can be used to bribe your way out of the country, or that you have a real liability in 1000 years, you may have a point. If the argument is that gold has done well in the past, sorry that is statistically insignificant nonsense that should be ignored.
GRP wrote: ↑Tue Aug 25, 2020 10:33 pm
AlohaJoe wrote: ↑Tue Aug 25, 2020 6:25 am
I understood what Uncorrelated posted. If you didn't, that doesn't make it word salad.
I understood what was said and I'm keen enough to know that the approach of using abstruse mathematics simply doesn't hold water in investing. (And you're talking to someone who otherwise LOVES math.)
Think about the massive statistical evidence behind, say, the value factor that culminated in the publishing of the famous 1992 Fama/French paper. You could hardly get more robust statistical results than those findings. Then what happened in the 27 year out of sample period? Value failed to deliver and practically disappeared.
It's not that I don't understand Uncorrelated's arguments. I do understand them and that's why I know they're baloney.
The value premium has been positive, see this paper: https://papers.ssrn.com/sol3/papers.cfm ... id=3525096
. However, the premium was too low to pass statistical significance tests (IIRC, the post-publication premium was 0.33 per month
and the t-stat was somewhere between 1 and 2). They say in the paper that they are unable to confirm that the premium still exists, but also unable to show that the premium was different than in the original paper. Furthermore, the value premium was strongly positive in international markets with a t-stat of 3.88 (europe), 4.73 (japan) and 3.98 (asia-pacific) in the time period 1992-2014. So no, the claim that value disappeared doesn't hold up to scrutiny. I don't see any statistical tests showing that the premium has disappeared.
Even if we suppose that what you're saying is true, it only means that we should demand far stronger evidence than provided in the 1992 Fama/French paper, and I agree with that. If I was an investor in 1992 I would be criticizing them for data mining. If your standard of proof is so high, there is no alternative to stocks and bonds. But if your standard of proof is so low that you're considering gold, then there are plenty of alternatives for you.
If you take a quantitative argument and omit the relevant statistical tests, that doesn't make it a qualitative arguments. It makes it a bad argument. Here's a hint: if you can show numbers or charts related to your argument, chances are low pretty low it's a qualitative argument.
I suggest you get your information elsewhere, this guy is only skilled at writing marketing material.