"...if some industry, for example semiconductors, collectively goes down 30% for some reason it is very unlikely that the typical investor will pick the one that somehow goes up."
So why doesn't that reasoning apply to countries? For example, there are 45 countries in the MSCI ACWI. If the market collectively goes down 30% for some reason, isn't it very unlikely that the typical investor will pick the one country that somehow goes up?
I addressed that many times before and then during the crisis.
My belief going in was that countries with different fundamental attributes (like commodity based) might be on different economic cycles which might mean they are on different stock market cycles such that there is some zig when the US zags. Also a fundamentally healthier country than the US might also only have a more cyclical event as opposed to secular.
Whether it was by luck or otherwise this is what happened and it happened with the countries I talked about before the crisis.
Brazil and Norway kept going up until June 2008 and Chile only endured a normal cyclical decline dropping a little over 30% as the US was going down 56% and Chile has since gone on to a new high.
I also wrote often about not expecting much real diversification from countries most similar to the US like Big Western Europe.
The work that lead me to these conclusions was far from complicated and involves a fair bit of common sense.
A reader (maybe you, I don't remember) was critical of my taking too simple of an approach to financials but to the extent a lot of the top down stuff I do relies on common sense means that it is doable for the typical investor.
I would add that this does not guarantee any sort of success but there is logic to be applied that has worked before. A country that is fundamentally healthy stands a far better chance for offering "normal" equity returns over some reasonably long period of time with the previous decade as exhibit A.
There is a lot of discussion right now about the extent to which correlations have gone up recently (here and here) but these are very short term observations. A hedge fund manager or the like probably does need to worry about these issues but you managing your own portfolio who presumably cares most about having enough money when you need it do not need to worry about short term correlation issues.