I noted something while reading this document. Very roughly the unemployment figure is 50% higher for males and twice as high for females. I was a first surprised by such a notable difference, but then I vaguely remembered that there are significantly more females working in the EU than in te USA (as for Japan, the breakdown isn’t given , but I even more vaguely remember that Japanese unemployment statistics have some quirk resulting in excluding many females from the unempoyment rate).
However I’m not very sure of it. Could someone confirm or deny that the the rate of women working is higher in the EU?
I’m affraid I can’t remember the way statistics are established in France. But I remember they have been modified twice in my living memory, each time resulting in a sudden drop of the number of people unemployed. The last reform, IIRC, was supposed to bring french statistics closer to the european/international norms (besides nicely reducing the unemployment figure) and if I’m not mistaken resulted in :
-excluding people not actively seeking a job, at least through official channels (the french unemployment agency. People will register with it in order to get unemployment benefits, but once these benefits stop, after two years, they might stop registering or use its services but still seek a job. They won’t be considered as unemployed anymore, though, IINM).
-excluding people having worked for any duration during the month, instead of only people who worked at least half-time.
Concerning differences, I also vaguely remember that US statistics are based on surveys while french ones are based on actual numbers based on people registered with the unemployed agency. But I’m not sure of this, either, ad anyway, it should’t make a significant difference if the survey is large enough and properly done.
I would also note that, for the same country, you could find different figures, one established according to the current rules in the country, and others by some international organization using its own methods. I just googled to find an information and my first result was article discussing a 1% discrepancy between the french official figure and the estimated figure published by a french economical survey organization using other criterias, supposedly closer to international norms (the official figure being at the time of the article 9.6 % and something like 8.6 % according to this organization). The article was too technical for me to understand (discussing the inclusion of lack thereof of “category 3” unemployed or such things) but apparently there were differences regarding people working in temporary agencies, people formerly self-employed, people working free-lance, etc…
I would also note, on the other hand, that people in “pre-retirment” (understand people who are in their 50s and instead of being just laid off, received an early and significantly lowered retirment package). In this case, the french figures would be under-estimated, since these presumably for the most part, accepted this not because they wouldn’t want a job but because they thought they couldn’t find one. And I assume that we could find other different instance of under-estimation. So, statistics are definitely tricky.
I would nevertheless tend to assume that estimated figures published by an international organization using its own criterias (rather than mentionning official figures from each country) would be fairly reliable as a comparison tool, even though they might underestimate unemployment everywhee by not including some categories. Of course, even in this case, there would be a bia if for instance an excluded category is more represented in some country than in another (for instance, for some reason, part-time jobs/ temporary jobs are more common in country X). Nevertheless, the difference in the figures available for the EU and the USA are so large that I strongly doubt they could be simply explained away by some statistical discrepancies.