On Tue, 2008-10-07 at 10:34 -0700, Jim Devine wrote: > Laurent answers: > > The street man definition of unemployment (100% minus employed divided > > by population) is perfectly observable. > > the population isn't really observable. Nor is employment.
Wow. Let me tell you a secret: nothing is observable! More seriously, each measure has a quality. Population and employment can be measured objectively at pretty good accuracy, something than can't be done with economist unemployment. > All we have > are guesses, though perhaps they are good ones. By the way, your > "street man" definition of unemployment is nothing but 100% minus the > employment/population ratio, a number that the US Bureau of Labor > Statistics regularly reports. Yep. Published doesn't mean it is used, isn't it? > To my mind, it's a mistake to look at any single number to understand > what's happening in the labor-power markets. One of the various > unemployment rates that the BLS reports, along with the E/POP ratio > and monthly payroll employment numbers seem to be a good combination. Yet unemployment is about the only measure used when discussing policy. We are in agreement here that more data is useful, here is what I said ""Payroll is observable, detailed distribution (sex/age/income per period) should be easy to collect and distribute, but of course economists prefer to keep this data secret just like they do for detailed price data and subjective quality adjustements."" sex/age/income distribution should be quite easy to get, but I've never seen it published. Am I missing a publication? > > The economist definition is, as the data mentionned in the rest of my > > post implies, a total joke and is not observable. > > Why is the official economist's definition a _total_ joke? the > question of who is without a job but is actively seeking one seems an > important question, even though it is not enough. (We should look at > involuntary part-timers, discouraged workers, etc. in addition.) > > By the way, the numbers on employment and those on unemployment have > similar problems and a similar degree of validity. Thus, the numbers > on the unemployment rate and your E/POP ratio have similar problems > and similar degrees of validity. That's completely wrong. population is 0/1 directly observable, employment is 0/1 and directly observable, "seeking a job hard enough to be counted as active unemployed" has probably more than a dozen possible values, is based on human psychology and not on facts so cannot be honnestly compared in accuracy with the two above measures. Your paragraph is in the category "total joke". > > BTW, how do you explain the data in my previous post? Number of > > economist paper studying this discrepancy is zero. Explain why? > > I remember looking at your data before and accepting the conclusion > that comparing unemployment rates between countries can be deceiving. > I don't have time right now for reading them again. Technical note: these are normalized numbers, OECD is supposed to make them comparable between countries. (Note 2: OECD hours worked number you see quoted everywhere are not normalized, and the OECD mention this fact but it is usually ignored). Even within a country, comparing unemployment at different point in time will be deceiving, see the NY Times graph here: http://guerby.org/blog/index.php/2008/03/09/179-le-new-york-times-lit-mon-blog With unemployment measure at same 4%, you have 12% of the USA 25-54 men population not working in 2008 whereas it was 9% in 1980: a one third difference. How good is the unemployment time serie for anything? Again, how do you explain HONNESTLY this fact together with the ones I mentionned initially? Laurent _______________________________________________ pen-l mailing list [email protected] https://lists.csuchico.edu/mailman/listinfo/pen-l
